Operation and maintenance process determination method and device

Abstract

The invention discloses an operation and maintenance process determination method and device, and relates to the technical field of computers. And the computing device determines alternative operation and maintenance flows associated with the operation and maintenance intention description information according to an operation and maintenance knowledge base containing operation and maintenance knowledge, and determines cue words according to the alternative operation and maintenance flows, so that the cue words contain the operation and maintenance knowledge associated with the operation and maintenance intention description information. Thus, the computing device can obtain the operation and maintenance knowledge associated with the operation and maintenance intention description information according to the cue word, training data is not needed to train an initial model to obtain the operation and maintenance knowledge, and the cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and particularly relates to an operation and maintenance process determination method and device. BACKGROUND

[0002] Operations refer to operation and maintenance work performed to ensure the stable operation of an information technology (IT) system. Generally, a computing device can train an initial artificial intelligence (AI) model (such as a large language model) using operation and maintenance professional knowledge as training data to obtain a trained model. The computing device can determine an operation and maintenance process based on an operation and maintenance intent input by a user using the trained model. The operation and maintenance process is used to indicate a process of calling an operation and maintenance tool to implement the operation and maintenance intent.

[0003] In the above manner, to ensure the matching of the operation and maintenance process and the operation and maintenance intent, a large amount of training data is required to train the initial AI model, which is costly. SUMMARY

[0004] The present application provides an operation and maintenance process determination method and device, which are used to solve the problem of high cost caused by training a large amount of training data to train an artificial intelligence model used to determine an operation and maintenance process.

[0005] In a first aspect, the present application provides an operation and maintenance process determination method. The method is executed by a computing device, and a first model is deployed on the computing device. The method includes: the computing device obtaining operation and maintenance intent description information input by a user. The computing device determines a candidate operation and maintenance process associated with the operation and maintenance intent description information based on an operation and maintenance knowledge base. The operation and maintenance knowledge base includes at least one operation and maintenance process. An operation and maintenance process is used to indicate operation and maintenance operations performed to implement an operation and maintenance intent and a calling manner of an operation and maintenance tool. The candidate operation and maintenance process is used to indicate operation and maintenance operations performed to implement an operation and maintenance intent corresponding to the operation and maintenance intent description information and a calling manner of an operation and maintenance tool. The computing device generates a prompt word according to the candidate operation and maintenance process and meta-information of an operation and maintenance tool in an operation and maintenance tool set. The operation and maintenance tool in the operation and maintenance tool set is determined according to the operation and maintenance intent description information. The computing device inputs the prompt word into the model to obtain a target operation and maintenance process. The target operation and maintenance process is used to indicate target operation and maintenance operations performed to implement the operation and maintenance intent description information and a calling manner of a target operation and maintenance tool.

[0006] If the computing device trains the initial model outputting the operation and maintenance process by using the training data, a large amount of training data is needed to obtain the trained model with operation and maintenance knowledge through a large amount of training. In the first aspect of the present application, the computing device determines the candidate operation and maintenance process associated with the operation and maintenance intent description information according to the operation and maintenance knowledge base containing operation and maintenance knowledge, and determines the prompt word according to the candidate operation and maintenance process, so that the prompt word contains the operation and maintenance knowledge associated with the operation and maintenance intent description information. In this way, the computing device can obtain the operation and maintenance knowledge associated with the operation and maintenance intent description information according to the prompt word, without training the initial model by using the training data to obtain the operation and maintenance knowledge, thereby reducing the cost.

[0007] In a possible implementation, the operation and maintenance knowledge base includes at least one operation and maintenance triple. One operation and maintenance triple includes an operation and maintenance intent, at least one first operation and maintenance operation for implementing the operation and maintenance intent and an operation sequence, and at least one first operation and maintenance tool for implementing the at least one first operation and maintenance operation and a calling mode of the at least one first operation and maintenance tool. One operation and maintenance triple corresponds to one operation and maintenance process. In this way, the operation and maintenance knowledge base includes at least one operation and maintenance triple corresponding to at least one operation and maintenance process. The operation and maintenance knowledge base provides rich operation and maintenance knowledge, which guarantees the computing device to obtain the candidate operation and maintenance process matching the operation and maintenance intent description information.

[0008] In another possible implementation, the target operation and maintenance process corresponds to a target operation and maintenance triple. The at least one operation and maintenance triple in the operation and maintenance knowledge base includes or does not include the target operation and maintenance triple. In this way, the computing device can directly determine the operation and maintenance triple corresponding to the candidate operation and maintenance process as the target operation and maintenance triple according to the actual application requirement, or adjust the operation and maintenance triple corresponding to the candidate operation and maintenance process and determine the adjustment result as the target operation and maintenance triple, thereby expanding the use scenario.

[0009] In a possible implementation, the model comprises: a first model and a second model. The computing device inputs the prompt word into the model to obtain the target operation and maintenance process, comprising: the computing device inputs the prompt word into the first model to generate operation and maintenance tool description information, operation and maintenance operation description information and result description information that implement the operation and maintenance intention description information. The computing device inputs the operation and maintenance tool description information, the operation and maintenance operation description information and the result description information into the second model to generate the calling mode of the target operation and maintenance tool. The second model is configured to: generate the calling mode of the target operation and maintenance tool according to a constraint decoding rule, the constraint decoding rule comprising: a plurality of operation and maintenance tools and at least one semantic label corresponding to the plurality of operation and maintenance tools. One semantic label corresponds to at least one of the plurality of operation and maintenance tools. One operation and maintenance tool corresponds to at least one of the at least one semantic label. The semantic label is used to indicate the application range of the operation and maintenance tool. In this way, the computing device cooperates with multiple models to obtain the target operation and maintenance process, different models implement different functions, the complexity of obtaining the target operation and maintenance process is reduced, the time required to obtain the target operation and maintenance process is shortened, the efficiency of obtaining the target operation and maintenance process is improved, and the accuracy of the data output by each model is improved.

[0010] In a possible implementation, before the computing device determines the candidate operation and maintenance process associated with the operation and maintenance intention description information based on the operation and maintenance knowledge base, the method further comprises: the computing device obtains at least one operation and maintenance record. One operation and maintenance record comprises: a second operation and maintenance operation, a tool identifier of a first operation and maintenance tool that implements the second operation and maintenance operation, and a result of calling the first operation and maintenance tool to execute the second operation and maintenance operation. The computing device generates an operation and maintenance process corresponding to the at least one operation and maintenance record according to the at least one operation and maintenance record. The operation and maintenance process corresponding to the at least one operation and maintenance record comprises: an operation and maintenance intention corresponding to the at least one operation and maintenance record, at least one second operation and maintenance operation and an operation sequence corresponding to the at least one operation and maintenance record, and at least one second operation and maintenance tool that implements the at least one second operation and maintenance operation and a calling mode of the at least one second operation and maintenance tool. The computing device constructs the operation and maintenance knowledge base according to the operation and maintenance process corresponding to the at least one operation and maintenance record. In this way, the computing device constructs the operation and maintenance knowledge base according to the at least one operation and maintenance record, converts the scattered and implicit operation and maintenance knowledge into structured, reusable and easily identifiable operation and maintenance knowledge for the model, and provides a guarantee for the operation and maintenance knowledge to be effectively used by the model.

[0011] In a possible implementation, the computing device determines the operation dependency relationship between the at least one second operation according to the at least one operation record, and determines the tool dependency relationship between the at least one second tool according to the at least one operation record. The computing device determines the operation intent corresponding to the at least one operation record according to the at least one second operation and the tool dependency relationship. The computing device determines the operation process corresponding to the at least one operation record according to the operation intent corresponding to the at least one operation record, the at least one second operation, the operation dependency relationship, the at least one second tool, and the tool dependency relationship. In this way, the computing device determines the operation process according to the operation intent corresponding to the at least one operation record, the at least one second operation, the operation dependency relationship, the at least one second tool, and the tool dependency relationship, which ensures that the determined operation process matches the operation intent, and further ensures the accuracy of the operation knowledge included in the operation knowledge base.

[0012] In a possible implementation, the computing device model includes a first model. The first model is configured to determine the operation intent according to the operation and the tool dependency relationship between the operation and a tool for implementing the operation. The computing device determines the operation intent corresponding to the at least one operation record according to the at least one second operation and the tool dependency relationship, including: determining, by the computing device, a dependency type of the tool dependency relationship. The dependency type includes one of linear, parallel, tree, directed acyclic graph, or combination. The computing device inputs the at least one second operation and the tool dependency relationship into the first model according to a prompt word template corresponding to the dependency type, to obtain the operation intent corresponding to the at least one operation record. In this way, the computing device inputs the at least one second operation and the tool dependency relationship into the first model according to the dependency type corresponding to the tool dependency relationship, which ensures that the first model can recognize the input information, and further ensures that the target operation process matches the operation intent description information.

[0013] In a possible implementation, the method further includes: in a case where the first information and the second information do not match, replacing, by the computing device, the first information indicated by the alternative operation process with the second information to obtain an updated operation knowledge base. The first information is: an operation and a calling manner of a tool for implementing an operation intent corresponding to operation intent description information. The second information is: an operation and a calling manner of a tool for implementing the operation intent description information. In this way, the computing device updates the operation knowledge base according to the target operation process of the operation description information to obtain an updated operation knowledge base, which ensures that the updated operation knowledge base matches the operation scenario.

[0014] In a possible implementation, before the computing device inputs the operation and maintenance tool description information, the operation and maintenance operation description information, and the result description information into the second model to generate the calling manner of the target operation and maintenance tool, the method further includes: determining, by the computing device, semantic labels corresponding to the plurality of operation and maintenance tools according to application scopes of the operation and maintenance tools. The target operation and maintenance tool belongs to the plurality of operation and maintenance tools. The computing device constructs constraint decoding rules based on the plurality of operation and maintenance tools and the semantic labels corresponding to the plurality of operation and maintenance tools. In this way, the computing device updates the semantic labels of the target operation and maintenance process, and provides a guarantee for matching the updated semantic labels with the operation and maintenance scene.

[0015] In a possible implementation, the method further includes: updating, by the computing device, the calling manner of the target operation and maintenance tool to the calling manner of the target operation and maintenance tool generated based on the constraint decoding rules. In this way, the computing device updates the calling manner of the target operation and maintenance tool according to the target operation and maintenance process, and provides a guarantee for matching the updated calling manner of the target operation and maintenance tool with the actual calling manner of the target operation and maintenance tool.

[0016] In a possible implementation, the first model is configured to: generate the operation and maintenance process according to the prompt word, and obtain the operation and maintenance intention according to a tool dependency relationship between the operation and maintenance operation and the operation and maintenance tool for implementing the operation and maintenance operation.

[0017] In a second aspect, the present application provides an operation and maintenance process determination apparatus. The apparatus includes various modules for performing the method in the first aspect or any possible design of the first aspect.

[0018] In a third aspect, the present application provides a processor. The processor includes an interface circuit and a control circuit. The interface circuit is configured to operate with the control circuit to implement the operation steps of the method in the first aspect or any possible design of the first aspect.

[0019] In a fourth aspect, the present application provides a computing device cluster. The computing device cluster includes at least one computing device, and each computing device includes a processor and a memory. The processor of the at least one computing device is configured to execute instructions stored in the at least one memory to cause the computing device cluster to perform the operation steps of the operation and maintenance process determination method in the first aspect or any possible design of the first aspect.

[0020] In a fifth aspect, the present application provides a computer-readable storage medium. The computer-readable storage medium includes: computer software instructions; and when the computer software instructions are executed in a computing device, the computer software instructions cause the computing device to perform the operation steps of the method in the first aspect or any possible implementation of the first aspect.

[0021] In a sixth aspect, the present application provides a computer program product. When the computer program product is run on a computer cluster, the computer program product causes the computer device cluster to perform the operation steps of the method as described in the first aspect or any possible implementation manner of the first aspect.

[0022] The beneficial effects of the above second aspect to sixth aspect can refer to the description of the first aspect or any implementation manner of the first aspect, which will not be repeated here. On the basis of the implementation manners of the above aspects provided by the present application, further combinations can be made to provide more implementation manners. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 Process schematic diagram for determining operation and maintenance process for large model;

[0024] Figure 2 Architecture schematic diagram of an operation and maintenance process determination system provided by the present application;

[0025] Figure 3 Structure schematic diagram of a chip provided by the present application;

[0026] Figure 4 Process schematic diagram of an operation and maintenance process determination method provided by the present application;

[0027] Figure 5 Process schematic diagram of constructing an operation and maintenance knowledge base provided by the present application;

[0028] Figure 6 Process schematic diagram of constructing constraint decoding rules provided by the present application;

[0029] Figure 7 Schematic block diagram of an operation and maintenance process determination method provided by the present application;

[0030] Figure 8 Schematic diagram of updating constraint decoding rules provided by the present application;

[0031] Figure 9 Process example diagram of an operation and maintenance process determination method provided by the present application;

[0032] Figure 10 Scene example diagram of an operation and maintenance process method provided by the present application;

[0033] Figure 11 Structure schematic diagram of an operation and maintenance process determination apparatus provided by the present application;

[0034] Figure 12 Structure schematic diagram of a computer device cluster provided by the present application;

[0035] Figure 13A connection diagram between computing devices is provided. DETAILED DESCRIPTION

[0036] For the sake of clear and concise description of the following embodiments, first give a brief introduction of related terms.

[0037] Artificial intelligence (ai) model: a mathematical model or computational framework that can simulate human intelligent behavior (such as reasoning, decision making, and recognition) by learning data through algorithms. Artificial intelligence models include but are not limited to large language models and the like. In some possible cases, artificial intelligence models can also be referred to as ai models, models, etc., which are not limited by the present application.

[0038] Large language model (llm): a generative deep learning model trained on massive training data (such as text data, image data, audio data, etc.). Large language models not only excel in understanding and generating human language, but also exhibit great potential in solving complex problems and tasks. In some possible examples, large language models can also be referred to as large models. The large model provided in the embodiments of the present application can not only refer to a large language model, but also refer to a model whose model parameters reach a certain degree. The large model can also refer to a model that contains various functions such as image processing function, human-computer interaction function, semantic search and dialogue function, etc. In this paper, in order to facilitate description, all are named as large model, but this should not be understood as a limitation of the present application, and will not be described in detail hereinafter.

[0039] Large models are widely used in the field of operation and maintenance. The large model determines the operation and maintenance process according to the operation and maintenance intention input by the user. Figure 1 The flowchart for determining the operation and maintenance process for the large model is shown in (a) of FIG. 1. Figure 1 The computing device can determine the operation and maintenance process by performing the following (1) to (3).

[0040] (1) The computing device obtains training data.

[0041] The computing device can collect original operation and maintenance data generated in the operation and maintenance scenario. The computing device generates extended operation and maintenance data according to the original operation and maintenance data. And the computing device constructs training data according to the original operation and maintenance data and the extended operation and maintenance data.

[0042] (2) The computing device trains a general large model using the training data to obtain a post-training large model.

[0043] (3) The computing device determines the operation and maintenance process according to the operation and maintenance intention input by the user by using the post-training large model.

[0044] In the process of large model inference (such as large model generation operation and maintenance process), the computing device can convert the regular expression into a finite state machine (fsm) to ensure that the large model outputs the operation and maintenance process in the specified format. For example Figure 1 As shown in (b) of Figure 1 As shown in (b) of Figure 1 As shown in (b) of

[0045] In the above process, in order to ensure that the operation and maintenance process matches the operation and maintenance intention, a large amount of training data is required to train the initial model, which is costly.

[0046] Therefore, the present application provides an operation and maintenance process determination method. In the method, the computing device determines a candidate operation and maintenance process associated with operation and maintenance intention description information according to an operation and maintenance knowledge base containing operation and maintenance knowledge, and determines a prompt word according to the candidate operation and maintenance process, so that the prompt word contains operation and maintenance knowledge associated with the operation and maintenance intention description information. In this way, the computing device can obtain operation and maintenance knowledge associated with the operation and maintenance intention description information according to the prompt word, without training an initial model with training data to obtain operation and maintenance knowledge, thereby reducing costs.

[0047] The above operation and maintenance process determination method can be applied to an operation and maintenance process determination system. Figure 2 The architecture of an operation and maintenance process determination system provided by the present application is shown in Figure 2 As shown in (b) of

[0048] The computing device 210 includes a communication interface 214, a processor 211, a memory 212, and a bus 216. The communication interface 214 is configured to communicate with devices located outside the computing device 210. For example, the computing device 210 receives user input data (such as operation and maintenance intention description information) through the communication interface 214. For another example, the computing device 210 sends a processing result (such as a target operation and maintenance process) to a user through the communication interface 214. The communication interface 214 can be an input / output (I / O) interface.

[0049] The processor 211 is the operation core and control core of the computing device 210, and can include a central processing unit (CPU), a specific integrated circuit, other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. In actual applications, the computing device 210 can also include multiple processors. The processor 211 can include one or more processor cores. The processor 211 is installed with an operating system and other software programs, so that the processor 211 can access the memory 212 and various peripheral component interconnect express (PCIe) devices. In some possible cases, at least one model, such as a first model, a second model, and the like, can be deployed on the processor 211. The first model and the second model can be the same type of model or different types of model, which is not limited in the present application. For example, the first model and the second model are both models for implementing interaction functions, such as large language models. According to actual application needs, the first model and the second model can have different parameter amounts, such as the parameter amount of the first model being greater than that of the second model, and the like.

[0050] The processor 211 is connected to the memory 212 through a bus 216. The bus 216 can be a double data rate (DDR) bus or other type of bus. The memory 212 is the main memory of the computing device 210. The memory 212 is usually used to store various running software in the operating system and the like. In order to improve the access speed of the processor 211, the memory 212 needs to have the advantage of fast access speed. In a conventional computer device, a dynamic random access memory (DRAM) is usually used as the memory 212. In addition to the DRAM, the memory 212 can also be other random access memories, such as a static random access memory (SRAM) and the like. In addition, the memory 212 can also be a read only memory (ROM). For the read only memory, for example, it can be a programmable read only memory (PROM), an erasable programmable read only memory (EPROM) and the like. The embodiment does not limit the number and type of the memory 212.

[0051] In some possible cases, the computing device 210 can also include a display 217. The computing device 210 displays the operation and maintenance intention input interface in the display 217. The computing device 210 obtains the operation and maintenance intention description information in response to the operation performed by the user on the operation and maintenance intention input interface. And the computing device 210 displays the target operation and maintenance process by using the display 217.

[0052] In some possible cases, in order to persistently store the data (such as the operation and maintenance intention description information, the target operation and maintenance process and the like), the operation and maintenance process determination system is also provided with a data storage system 213. The data storage system 213 can be located outside the computing device 210 (such as shown in the figure), and exchanges data with the computing device 210 through a network. In some possible scenarios, the data storage system 213 can also be located inside the host, such as the data storage system 213 exchanges data with the processor 211 through the bus 216. At this time, the data storage system 213 can be a hard disk. Figure 2

[0053] The processor 211 in the example shown in the figure can be implemented by a chip, such as shown in the figure, the processor 211 can be implemented by a chip. Figure 2 Figure 3 Figure 3 ​​​A structural schematic diagram of a chip is provided in the present application. As an example, the chip 300 includes a core 301, a CPU 302, a system buffer 303, an input / output (I / O) device 305, and a DDR 306.

[0054] The CPU 302 is configured to accept a task (such as an operation and maintenance process determination task) and call the core 301 to execute the task. In the case where the chip 300 has multiple cores 301, the CPU 302 is also configured to undertake a scheduled task. For example, the CPU 302 can be implemented by an ARM processor, which is small in size, low in power consumption, adopts a 34-bit reduced instruction set, and is simple and flexible in addressing. Of course, in some embodiments, the CPU 302 can also be implemented by other processors.

[0055] The core 301 is configured to provide the computing power required to obtain an operation and maintenance process determination task. In an optional case, the core 301 includes a load / store unit (LSU), a cube computing unit, a scalar computing unit, a vector computing unit, and a buffer. The LSU is configured to load data to be processed and store processed data, and can also be used for read / write management of internal data in the core between different buffers, and for completing some format conversion operations. The cube computing unit is configured to provide the core computing power of matrix multiplication. The scalar computing unit is a single instruction single data (SISD) processor, which processes only one piece of data (usually an integer or a floating point number) at the same time. The vector computing unit, also known as an array processor, is a processor that can directly operate a group of arrays or vectors for computation. The number of buffers can be one or more, such as a level 1 buffer (L1 buffer). The buffer is used to temporarily store some data that needs to be repeatedly used by the core 301, so as to reduce read / write from the bus. In addition, the implementation of some data format conversion functions also requires that the source data be located in the buffer.

[0056] The system buffer 303, mainly a level 2 buffer, is configured to temporarily store input data, intermediate results, or final results of the chip.

[0057] The DDR 306 is an off-chip memory, which can also be replaced by a high bandwidth memory (HBM) or other off-chip memory. The DDR 306 is located between the chip and the external memory, and overcomes the access speed limitation when the computing resource shares the memory for read / write.

[0058] The I / O device 305 included in the chip 300 refers to hardware for data transmission, and can also be understood as a device connected with the I / O interface. Common I / O devices include network cards, printers, keyboards, mice, etc. All external storages can also be used as I / O devices, such as hard disks, floppy disks, optical disks, etc.

[0059] The core 301, the CPU 302, the system buffer 303, the I / O device 305, and the DDR 306 are connected through a bus. The bus can include a channel for transmitting information between the above components (such as the CPU 302 and the system buffer 303). In addition to the data bus, the bus can also include a power bus, a control bus, and a state signal bus, etc. However, for the purpose of clear illustration, the bus can be a PCIe bus, or an extended industry standard architecture (EISA) bus, a unified bus (Ubus or UB), a compute express link (CXL), a cache coherent interconnect for accelerators (CCIX), etc. For example, the core 301 can access the I / O device 305 through the PCIe bus. The core 301 is connected with the system buffer 303 through the DDR bus. Here, different system buffers 303 can use different data buses to communicate with the core 301, and therefore, the DDR bus can also be replaced by other types of data buses, and the bus type is not limited in the embodiments of the present application.

[0060] For example, after the CPU 302 loads the data (such as operation and maintenance intention description information) to be processed by the operation and maintenance process determination task into the DDR 306, the LSU in the core 301 reads the data from the DDR 306, processes the data to obtain a processing result (such as a target operation and maintenance process), and then stores the processing result in the DDR 306. In the case where the operation and maintenance process determination system 200 includes a data storage system 213, the processing result can be sent to the data storage system 213 by the network interface card for persistent storage.

[0061] It can be understood that the structure illustrated in the embodiments does not constitute a specific limitation on the computing device and the chip. In other embodiments, the computing device and the chip can include more or fewer components than those illustrated, or combine certain components, or split certain components, or different arrangement of components. The illustrated components can be implemented in hardware, software, or a combination of software and hardware.

[0062] The following will be described in combination with Figure 2 andFigure 3 The content shown, the operation and maintenance process determination method provided in the present application is described in detail.

[0063] Figure 4 The flowchart of the operation and maintenance process determination method provided in the present application, the operation and maintenance process determination method can be executed by a single computing device, a computing device cluster including a plurality of computing devices, a component of a computing device (such as a processor, a chip or a chip system of a computing device), etc., and can also be implemented by a logic module or software. In the case of executing the operation and maintenance process determination method by a single computing device, the computing device can be Figure 2 The computing device 210 shown, for the hardware implementation of the computing device, please refer to the foregoing Figure 2 The description, here will not be repeated. In the case of executing the operation and maintenance process determination method by a component of a computing device, the computing device component can be Figure 3 The chip shown, for the hardware implementation of the chip, please refer to the foregoing Figure 3 The description, here will not be repeated. In some optional examples, the operation and maintenance process determination method can also be executed by other computing devices, for the hardware implementation of the computing device, please refer to the foregoing Figure 2 And Figure 3 The description, here will not be repeated.

[0064] Here, a single computing device executes the method provided in the present embodiment, and the computing device deployment model is taken as an example for description, as Figure 4 The operation and maintenance process determination method provided in the present embodiment includes the following S410 to S440.

[0065] S410, the computing device acquires the operation and maintenance intention description information input by the user.

[0066] The operation and maintenance intention description information is used to indicate the operation and maintenance intention of the user. The presentation form of the operation and maintenance intention description information includes but is not limited to at least one of the following: text, image, audio, video, etc.

[0067] The computing device can acquire the operation and maintenance intention description information in various ways, two possible ways are given as follows.

[0068] Example 1, the computing device acquires the operation and maintenance intention description information based on the operation of the user.

[0069] In this case, the computing device can provide an operation and maintenance intention input interface, and obtain operation and maintenance intention description information in response to user operations on the operation and maintenance intention input interface. The user operations on the operation and maintenance intention input interface include, but are not limited to, input operations, click operations, and the like. For example, the computing device provides an operation and maintenance intention input interface. The operation and maintenance intention input interface includes an operation and maintenance intention input window. In response to the user inputting "Help me analyze the commonalities and abnormalities of these hosts, and the host information is as follows: the Internet protocol (IP) address of host 1 is ip 1 to the IP address of host n is ip n" in the operation and maintenance intention input window, the computing device obtains operation and maintenance intention description information as follows: "Analyze the commonalities and abnormalities between host 1 and host n, and the IP addresses of host 1 to host n are ip 1 to ip n".

[0070] In Example 2, the computing device receives operation and maintenance intention description information input by a user using another device.

[0071] The other device can be a device communicatively connected to the computing device, including but not limited to a personal computer, a mobile phone, a USB flash disk, a hard disk, and the like. In this case, the computing device can receive operation and maintenance intention description information transmitted by the other device.

[0072] The above describes two possible ways for the computing device to obtain operation and maintenance intention description information. Depending on actual application needs, the computing device can also obtain operation and maintenance intention description information in other ways, which are not limited in the present application.

[0073] Compared with directly generating an operation and maintenance process according to the operation and maintenance intention description information using the first model, after obtaining the operation and maintenance intention description information, the computing device can also use existing operation and maintenance knowledge to enhance the operation and maintenance intention description information, generate prompt words containing sufficient operation and maintenance knowledge based on the enhanced operation and maintenance intention description information, and use the first model to obtain a target operation and maintenance process matching the operation and maintenance intention description information based on the prompt words. In this way, the first model can obtain sufficient operation and maintenance professional knowledge, which guarantees that the target operation and maintenance process generated by the first model based on the prompt words matches the operation and maintenance intention description information. Without using a large amount of training data to train the initial model to have sufficient operation and maintenance knowledge, and then guaranteeing that the generated target operation and maintenance process matches the operation and maintenance intention description information, costs are saved.

[0074] The computing device can perform the following S420 and S430 to make the prompt words contain sufficient operation and maintenance knowledge.

[0075] S420, the computing device determines a candidate operation and maintenance process associated with the operation and maintenance intention description information based on an operation and maintenance knowledge base.

[0076] The operation and maintenance knowledge base comprises at least one operation and maintenance process. An operation and maintenance process is used to indicate operation and maintenance operations performed to achieve an operation and maintenance intention and a calling manner of operation and maintenance tools. The alternative operation and maintenance process is used to indicate operation and maintenance operations performed to achieve an operation and maintenance intention corresponding to operation and maintenance intention description information and a calling manner of operation and maintenance tools.

[0077] The operation and maintenance knowledge base is used to indicate operation and maintenance execution. The operation and maintenance knowledge base comprises at least one operation and maintenance triple. An operation and maintenance triple comprises an operation and maintenance intention, at least one first operation and maintenance operation and operation sequence for achieving the operation and maintenance intention, and at least one first operation and maintenance tool and a calling manner of the at least one first operation and maintenance tool for achieving the at least one first operation and maintenance operation. An operation and maintenance triple corresponds to an operation and maintenance process. In some possible examples, an operation and maintenance tool can also be referred to as an operation and maintenance tool type, and an operation and maintenance operation can also be referred to as an operation and maintenance operation type.

[0078] In some possible cases, one operation and maintenance tool can implement one or more operation and maintenance operations, and one operation and maintenance operation can be implemented by one or more operation and maintenance tools, which are not limited in the present application. For example, operation and maintenance tool 1 supports implementing operation and maintenance operation 1, and operation and maintenance tool 2 supports implementing operation and maintenance operation 2 and operation and maintenance operation 3. For another example, operation and maintenance operation 4 is implemented by operation and maintenance tool 3, and operation and maintenance operation 5 is implemented by operation and maintenance tool 4 and operation and maintenance tool 5. The following describes the method provided by the present application by taking one operation and maintenance tool implementing one operation and maintenance operation and one operation and maintenance operation being implemented by one operation and maintenance tool as examples.

[0079] For example, the operation and maintenance knowledge base comprises operation and maintenance triple 1. The operation and maintenance triple 1 is <batch host commonality and exception analysis, query host i location, query host i status, query host i packet loss, query host i associated switch, query host i switch alarm, query host i switch change, tool 1, tool 2, tool 3, tool 4, tool 5, tool 6>. In this case, the "batch host commonality and exception analysis" in the operation and maintenance triple 1 is an operation and maintenance intention. The "query host i location, query host i status, query host i packet loss, query host i associated switch, query host i switch alarm, query host i switch change" in the operation and maintenance triple 1 are six first operation and maintenance operations and operation sequences for achieving the operation and maintenance intention "batch host commonality and exception analysis". The "tool 1, tool 2, tool 3, tool 4, tool 5, tool 6" in the operation and maintenance triple 1 are six operation and maintenance tools for implementing the six first operation and maintenance operations and a calling sequence of the six operation and maintenance tools.

[0080] The operation and maintenance knowledge base is described above by taking examples. For a process of constructing the operation and maintenance knowledge base by the computing device, refer to related descriptions below, which are not described herein again. The process of determining the alternative operation and maintenance process by the computing device is described below.

[0081] The computing device can determine the candidate operation and maintenance process according to semantic similarity between the user input operation and maintenance intention description information and a similarity element in at least one operation and maintenance triple included in the operation and maintenance knowledge base. The similarity element includes at least one of the following: the operation and maintenance intention, and at least one first operation and maintenance operation.

[0082] Specifically, the computing device can perform the following (1) and (2) to determine the candidate operation and maintenance process.

[0083] (1) The computing device obtains at least one similarity according to the user input operation and maintenance intention description information and at least one similarity element in at least one operation and maintenance triple of the operation and maintenance knowledge base.

[0084] The similarity is the similarity between the user input operation and maintenance intention description information and one of the at least one operation and maintenance triple of the operation and maintenance knowledge base.

[0085] The following takes the computing device obtaining a first similarity between the user input operation and maintenance intention description information and a first operation and maintenance triple of the operation and maintenance knowledge base as an example to describe the process of the computing device obtaining the similarity.

[0086] According to different similarity elements, the computing device determines the first similarity in different ways, which are described as follows.

[0087] Case 1: The similarity element is the operation and maintenance intention in the first operation and maintenance triple.

[0088] In this case, the computing device obtains semantic similarity between the user input operation and maintenance intention description information and the operation and maintenance intention in the first operation and maintenance triple, and the computing device determines the first similarity according to the semantic similarity. The computing device can directly take the semantic similarity as the first similarity, or perform an operation on the semantic similarity and take the operation result as the first similarity, which is not limited in the present application.

[0089] Case 2: The similarity element is at least one first operation and maintenance operation in the first operation and maintenance triple.

[0090] In this case, the computing device obtains at least one semantic similarity between the user input operation and maintenance intention description information and at least one first operation and maintenance operation in the first operation and maintenance triple, and the computing device determines the first similarity according to the at least one semantic similarity. The computing device can determine the first similarity according to the at least one semantic similarity in various ways, including but not limited to: taking the average of the at least one semantic similarity as the first similarity, taking the sum of the at least one semantic similarity as the first similarity, taking the maximum of the at least one semantic similarity as the first similarity, and the like.

[0091] Case 3, the similarity element is the operation and maintenance intention in the first operation and maintenance triple and at least one first operation and maintenance operation.

[0092] In this case, the computing device can obtain the operation and maintenance intention description information input by the user and the semantic similarity of the operation and maintenance intention in the first operation and maintenance triple, the at least one semantic similarity between the operation and maintenance intention description information input by the user and the at least one first operation and maintenance operation in the first operation and maintenance triple, and the first similarity according to the semantic similarity corresponding to the operation and maintenance intention and the at least one semantic similarity corresponding to the at least one first operation and maintenance operation in the manner described in the above case 1 and case 2. Similarly, the computing device can obtain the first similarity according to the semantic similarity corresponding to the operation and maintenance intention and the at least one semantic similarity corresponding to the at least one first operation and maintenance operation in the manner described above. For more details about the similarity determination manner, please refer to the above description, which will not be repeated here.

[0093] (2) The computing device determines the operation and maintenance process corresponding to the one or more operation and maintenance triples whose similarity meets the similarity condition as the candidate operation and maintenance process.

[0094] The similarity condition includes but is not limited to: the similarity is greater than or equal to the similarity threshold. The similarity threshold can be equal to the similarity corresponding to the operation and maintenance triple located at a specified order in the similarity sequence. The similarity sequence is a sequence obtained by sorting at least one operation and maintenance triple in the operation and maintenance knowledge base in descending or ascending order of similarity. The specified order can be preset or set according to the actual application, which is not limited in the present application. For example, the similarity sequence is a sequence obtained by sorting at least one operation and maintenance triple in the operation and maintenance knowledge base in descending order of similarity. The specified order is 1. In this case, the computing device selects the operation and maintenance triple with the largest similarity between the operation and maintenance intention description information input by the user from at least one operation and maintenance triple in the operation and maintenance knowledge base, and determines the operation and maintenance process corresponding to the operation and maintenance triple with the largest similarity as the candidate operation and maintenance process.

[0095] The above describes the process of determining the candidate operation and maintenance process by the computing device taking the candidate operation and maintenance process as one operation and maintenance process. According to the actual application, the computing device can also determine multiple operation and maintenance processes as candidate operation and maintenance processes, which is not limited in the present application. The operation and maintenance process determination method provided by the present application will be described below taking the candidate operation and maintenance process as one operation and maintenance process.

[0096] After performing S420 to determine the candidate operation and maintenance process, the computing device can also perform the following S430 to generate a prompt word.

[0097] S430, the computing device generates a prompt word according to the candidate operation and maintenance process and the meta information of the operation and maintenance tool in the operation and maintenance tool set.

[0098] Among them, the operation and maintenance tools in the operation and maintenance tool set are determined according to the operation and maintenance intention description information.

[0099] According to different actual application scenarios, the contents included in the operation and maintenance tool set are different, which are not limited in the present application.

[0100] For example, the operation and maintenance tool set can include all operation and maintenance tools.

[0101] In this case, the computing device can determine all operation and maintenance tools as operation and maintenance tools in the operation and maintenance tool set.

[0102] In some possible cases, the computing device can determine the operation and maintenance tools corresponding to at least one operation and maintenance triple in the operation and maintenance knowledge base as all operation and maintenance tools. According to the needs of actual application, the computing device can also determine part of operation and maintenance tools corresponding to the operation and maintenance knowledge base as all operation and maintenance tools, and the computing device can also determine other operation and maintenance tools not included in the operation and maintenance knowledge base as all operation and maintenance tools, which are not limited in the present application.

[0103] For another example, the operation and maintenance tool set can include operation and maintenance tools determined according to the operation and maintenance intention description information. The process of determining operation and maintenance tools included in the operation and maintenance tool set by the computing device is described as follows.

[0104] In this case, the computing device can determine operation and maintenance tools in the operation and maintenance tool set according to the operation and maintenance intention description information, or determine operation and maintenance tools in the operation and maintenance tool set according to the alternative operation and maintenance process, which are not limited in the present application.

[0105] In the case that the computing device determines operation and maintenance tools in the operation and maintenance tool set according to the operation and maintenance intention description information, the computing device can determine operation and maintenance tools in the operation and maintenance tool set by using the following process. Specifically, the computing device selects operation and maintenance tools with a similarity meeting a requirement from all operation and maintenance tools, and determines the operation and maintenance tools with the similarity meeting the requirement as operation and maintenance tools in the operation and maintenance tool set. The operation and maintenance tools in the operation and maintenance tool set can be operation and maintenance tools with a semantic similarity meeting a semantic similarity condition from all operation and maintenance tools.

[0106] In the case that the computing device determines operation and maintenance tools in the operation and maintenance tool set according to the alternative operation and maintenance process, the computing device can determine operation and maintenance tools in the operation and maintenance tool set by using the following process. Specifically, the computing device selects operation and maintenance tools with a similarity meeting a requirement from all operation and maintenance tools, and determines the operation and maintenance tools with the similarity meeting the requirement as operation and maintenance tools in the operation and maintenance tool set. The operation and maintenance tools in the operation and maintenance tool set can be operation and maintenance tools with a semantic similarity meeting a semantic similarity condition from all operation and maintenance tools.

[0107] In some possible scenarios, the set of operation and maintenance tools can include all operation and maintenance tools corresponding to the alternative operation and maintenance process, and can further include more or less operation and maintenance tools, which are not limited in the present application.

[0108] In some possible scenarios, the computing device can further determine the number of operation and maintenance tools included in the set of operation and maintenance tools according to the number of tokens supported by the large model input, and can further determine the number of operation and maintenance tools included in the set of operation and maintenance tools according to other manners, which are not limited in the present application.

[0109] The operation and maintenance tools are described above, and the way of generating the prompt is described below.

[0110] Specifically, the computing device can determine operation and maintenance triples corresponding to the alternative operation and maintenance process. The computing device obtains an operation and maintenance intent, at least one first operation and maintenance tool for implementing the operation and maintenance intent, and a calling sequence of the at least one first operation and maintenance tool in the operation and maintenance triples corresponding to the alternative operation and maintenance process. The computing device determines a set of operation and maintenance tools according to the operation and maintenance intent description information. The set of operation and maintenance tools includes tool identifiers of at least one operation and maintenance tool. The computing device generates a prompt according to the operation and maintenance intent corresponding to the alternative operation and maintenance process, the at least one first operation and maintenance tool corresponding to the alternative operation and maintenance process, the calling sequence of the at least one first operation and maintenance tool, and the meta information of the operation and maintenance tools in the set of operation and maintenance tools. In some possible examples, the prompt can also be referred to as prompt.

[0111] The process of generating the prompt by the computing device is exemplarily described below.

[0112] In this example, the at least one operation and maintenance three-dimensional group in the operation and maintenance knowledge base corresponds to operation and maintenance tool 1 to operation and maintenance tool 10. The operation and maintenance intention description information input by the user is "help me analyze what commonalities and abnormalities these hosts have, and the host information is as follows: the Internet protocol address ip 1 of host 1 to the ip n of host n". The alternative operation and maintenance process is operation and maintenance process 1, and operation and maintenance process 1 corresponds to operation and maintenance three tuple 1. Operation and maintenance three tuple 1 is <batch host commonality and abnormality analysis, query host i host location-query host i host state-query host i host packet loss-query host i associated switch-query host i switch alarm-query host i switch change, tool 1-tool 2-tool 3-tool 4-tool 5-tool 6>. In this case, the computing device can generate a prompt word according to the meta information of operation and maintenance tool 1 to operation and maintenance tool 10 and operation and maintenance process 1. The prompt word is "the current operation and maintenance tool set includes: the meta information of operation and maintenance tool 1 to operation and maintenance tool 6. The similar problem example is <batch host commonality and abnormality analysis, query host i host location-query host i host state-query host i host packet loss-query host i associated switch-query host i switch alarm-query host i switch change, tool 1-tool 2-tool 3-tool 4-tool 5-tool 6>, based on the above information, complete the user's request as follows: analyze what commonalities and abnormalities the following hosts have, and the host information is as follows: the ip 1 to ip n of host 1 to host n".

[0113] The above describes the process of generating a prompt word using S430, and the process of obtaining a target operation and maintenance process is described below in combination with S440.

[0114] S440, the computing device inputs the prompt word into the model to obtain a target operation and maintenance process.

[0115] The target operation and maintenance process is used to indicate that the target operation and maintenance operation corresponding to the operation and maintenance intention description information is implemented and the calling manner of the target operation and maintenance tool is implemented.

[0116] In some possible cases, implementing the target operation and maintenance operation corresponding to the operation and maintenance intention description information includes implementing the target operation and maintenance operation and the operation order of executing the target operation and maintenance operation. The calling manner of the target operation and maintenance tool includes the target operation and maintenance tool that implements the target operation and maintenance operation, the order of calling the target operation and maintenance tool, and the calling specification of calling the target operation and maintenance tool. The calling specification of the target operation and maintenance tool is used to indicate the calling rule of calling the target operation and maintenance tool. The calling specification of the target operation and maintenance tool includes but is not limited to the argument corresponding to the target operation and maintenance tool and the actual argument corresponding to the target operation and maintenance tool.

[0117] In some possible scenarios, the target operation and maintenance process corresponds to a target operation and maintenance triple. At least one operation and maintenance triple in the operation and maintenance knowledge base includes or does not include the target operation and maintenance triple. That is, the target operation and maintenance process can be the same as the alternative operation and maintenance process, or can be an operation and maintenance process different from the alternative operation and maintenance process generated according to the alternative operation and maintenance process. The model can be a model that implements an interactive function, including but not limited to a large language model. The model is used to generate an operation and maintenance process according to a prompt word.

[0118] In some possible scenarios, the model can be one model or multiple models, which are not limited in the present application. When the model includes multiple models, the multiple models can be the same model or different models, which are not limited in the present application. When the multiple models are the same model, the multiple models can have the same parameter amount or different parameter amounts, which are not limited in the present application. For example, the model includes a first model and a second model. The first model and the second model are both large language models, and the first model includes more model parameters than the second model. The first model is used to generate operation and maintenance tool description information, operation and maintenance operation description information, and result description information that implement operation and maintenance intent description information according to a prompt word. The second model is used to generate a calling manner of the operation and maintenance tool according to a constraint decoding rule. In some possible examples, the first model can also be referred to as a large llm or a brain, and the second model can also be referred to as a small llm or a brain.

[0119] The following describes a process of obtaining a target operation and maintenance process by taking a first model and a second model used by a computing device as an example. Specifically, the process can include the following (1) and (2).

[0120] (1) The computing device inputs a prompt word into the first model to generate operation and maintenance tool description information, operation and maintenance operation description information, and result description information that implement operation and maintenance intent description information.

[0121] The operation and maintenance tool description information is used to indicate an operation and maintenance tool. The operation and maintenance operation description information is used to indicate an operation and maintenance operation. The result description information is used to indicate a result of calling the operation and maintenance tool to perform the operation and maintenance operation. For example, the computing device inputs a prompt word 1 into the first model to generate a first step to a third step that implement operation and maintenance description information. The first step includes performing an operation 1 (such as querying a host location) by using an operation and maintenance tool (such as an operation and maintenance tool 6) to obtain a result 1 (such as a host location). The second step includes performing an operation 2 (such as querying a host status) by using an operation and maintenance tool (such as an operation and maintenance tool 7) to obtain a result 2 (such as a host status). The third step includes performing an operation 3 (such as querying a host packet loss) by using an operation and maintenance tool 3 (such as an operation and maintenance tool 9) to obtain a result 3 (such as a host packet loss). The first step to the third step are operation and maintenance tool description information, operation and maintenance operation description information, and result description information.

[0122] In some possible scenarios, after the computing device inputs the prompt word into the first model, the first model can further generate analysis description information. The analysis description information is used to indicate an analysis process for implementing the operation and maintenance intention description information determined by the computing device according to the prompt word.

[0123] (2) The computing device inputs the operation and maintenance tool description information, the operation and maintenance operation description information, and the result description information into the second model to generate a calling manner of the target operation and maintenance tool.

[0124] The second model is configured to generate the calling manner of the target operation and maintenance tool according to a constraint decoding rule. The constraint decoding rule includes a plurality of operation and maintenance tools and at least one semantic label corresponding to the plurality of operation and maintenance tools. One semantic label corresponds to at least one of the plurality of operation and maintenance tools. One operation and maintenance tool corresponds to at least one of the at least one semantic label. The semantic label is used to indicate an application range of the operation and maintenance tool.

[0125] In some possible scenarios, the constraint decoding rule is a rule in which the model determines a next output according to a current output. The constraint decoding rule includes a tool label constraint rule, a tool parameter constraint rule, and a parameter argument constraint rule. The constraint decoding rule includes various contents, which are described below by using (i) to (iii).

[0126] (i) The tool label constraint rule.

[0127] The tool label constraint rule is used to indicate a constraint relationship between the semantic label and the operation and maintenance tool. The tool label constraint rule includes a plurality of operation and maintenance tools and at least one semantic label corresponding to the plurality of operation and maintenance tools. One semantic label corresponds to at least one of the plurality of operation and maintenance tools. One operation and maintenance tool corresponds to at least one of the at least one semantic label. For example, the plurality of operation and maintenance tools include operation and maintenance tools 1 to operation and maintenance tools n. The operation and maintenance tools corresponding to the semantic label 1 in the operation and maintenance tools 1 to operation and maintenance tools n include operation and maintenance tools 1 to operation and maintenance tools 4.

[0128] In some possible scenarios, the semantic label can be determined by using a label determination manner, which includes but is not limited to being determined according to a function implemented by the operation and maintenance tool, being determined according to an operation and maintenance intention, being determined according to a fault scenario, and the like. In a case where the semantic label is determined according to the function implemented by the operation and maintenance tool, the application range of the operation and maintenance tool indicated by the semantic label is a range corresponding to the function implemented by the operation and maintenance tool. In a case where the semantic label is determined according to the operation and maintenance intention, the application range of the operation and maintenance tool indicated by the semantic label is a range corresponding to the operation and maintenance intention. In a case where the semantic label is determined according to the fault scenario, the application range of the operation and maintenance tool indicated by the semantic label is a range corresponding to the fault scenario.

[0129] Table 1 is an example of the correspondence between semantic labels and operation and maintenance tools

[0130] Semantic label Tool identification of operation and maintenance tool Alarm Operation and maintenance tool 1 Change Operation and maintenance tool 2 Whole line topology Operation and maintenance tool 3 Monitoring index Operation and maintenance tool 4 Customer fault reporting Operation and maintenance tool 5 In-band state / out-of-band state Operation and maintenance tool 6 Host ping packet loss Operation and maintenance tool 7 Monitoring index Operation and maintenance tool 8 Switch port state Operation and maintenance tool 7 Host packet loss Operation and maintenance tool 9 Query Operation and maintenance tool 6 Host state Operation and maintenance tool 7

[0131] In some possible scenarios, the computing device can determine semantic labels of the plurality of operation and maintenance tools according to application scopes of the operation and maintenance tools, and construct constraint decoding rules based on the plurality of operation and maintenance tools and the semantic labels of the plurality of operation and maintenance tools.

[0132] In some possible scenarios, the computing device can further generate operation and maintenance tool description information implementing the operation and maintenance intent description information based on the constraint decoding rules. Specifically, the computing device inputs the prompt word into the first model, and the first model generates the operation and maintenance tool description information implementing the operation and maintenance intent description information based on the constraint decoding rules. In this way, the matching degree between the operation and maintenance tool description information generated by the first model and the operation and maintenance intent description information is improved.

[0133] (ii) Tool parameter constraint rules.

[0134] The tool label constraint rules are used to indicate constraint relationships between operation and maintenance tools and parameters. The tool label constraint rules include at least one parameter corresponding to an operation and maintenance tool. For example, the plurality of operation and maintenance tools includes operation and maintenance tool 1. The operation and maintenance tool 1 has parameter 1 and parameter 2.

[0135] (iii) Parameter argument constraint rules.

[0136] The parameter argument constraint rules are used to indicate constraint relationships between parameters and arguments. The parameter argument constraint rules include parameters of an operation and maintenance tool and arguments corresponding to the parameters of the operation and maintenance tool. For example, the plurality of operation and maintenance tools includes operation and maintenance tool 1. The operation and maintenance tool 1 has parameter 1 and parameter 2. The parameter 1 is an IP address, and the parameter 2 is a host location. In this case, the argument corresponding to the parameter 1 can be 192.168.1.x. The argument corresponding to the parameter 2 can be city 1.

[0137] The constraint decoding rules are described above. For the process of constructing the constraint decoding rules by the computing device, refer to the related description below Figure 6 , which is not described herein again.

[0138] The process of the computing device generating a calling mode of the target operation and maintenance tool by using the second model is described below.

[0139] The computing device inputs the operation and maintenance tool description information, the operation and maintenance operation description information, and the result description information into the second model. The second model can select a semantic label matching the operation and maintenance operation description information from at least one semantic label. The second model further determines a second operation and maintenance tool and a calling specification of the second operation and maintenance tool from the operation and maintenance tools corresponding to the semantic label according to the result description information.

[0140] Specifically, the computing device inputs the operation and maintenance tool description information, the operation and maintenance operation description information, and the result description information into the second model. The computing device obtains at least one first probability between the operation and maintenance operation description information and at least one semantic label by using the second model. The computing device selects a target first probability satisfying a condition (such as the largest first probability) from the at least one first probability by using the second model, and determines that a semantic label corresponding to the target first probability is a target semantic label matched with the operation and maintenance operation description information. The computing device obtains at least one second probability between the result description information and at least one operation and maintenance tool corresponding to the target semantic label by using the second model. The computing device selects a target second probability satisfying a condition (such as the largest second probability) from the at least one second probability by using the second model, and determines that an operation and maintenance tool corresponding to the target second probability is a target operation and maintenance tool. The computing device obtains at least one third probability between the target operation and maintenance tool and at least one parameter. The computing device selects a target third probability satisfying a condition (such as the largest two third probabilities) from the at least one third probability by using the second model, and determines that a parameter corresponding to the target third probability is a parameter of the target operation and maintenance tool. The computing device obtains at least one fourth probability between the parameter and at least one argument. The computing device selects a target fourth probability satisfying a condition (such as the largest two fourth probabilities) from the at least one fourth probability by using the second model, and determines that an argument corresponding to the target fourth probability is an argument corresponding to the parameter. The computing device generates a calling specification of the target operation and maintenance tool according to the parameter and the argument of the target operation and maintenance tool.

[0141] In some possible cases, the operation and maintenance tool indicated by the operation and maintenance tool description information generated by the first model can be the same as the second operation and maintenance tool determined by the second model, and the operation and maintenance tool indicated by the operation and maintenance tool description information generated by the first model can also be different from the second operation and maintenance tool determined by the second model, which is not limited in the present application. In the case that the operation and maintenance tool indicated by the operation and maintenance tool description information generated by the first model is different from the second operation and maintenance tool determined by the second model, the computing device determines the second operation and maintenance tool determined by the second model as a target operation and maintenance tool for implementing the operation and maintenance intention description information, and determines a calling specification of the second operation and maintenance tool as a calling specification of the target operation and maintenance tool.

[0142] For example, the computing device uses a first model to generate operation and maintenance tool description information, operation and maintenance operation description information, and result description information that implement the operation and maintenance intent description information, including: calling operation and maintenance tool 1 to implement operation and maintenance operation 1 and obtaining result 1. In this case, the computing device inputs "calling operation and maintenance tool 1 to implement operation and maintenance operation 1 and obtaining result 1" into the second model. The computing device uses the second model to determine semantic label 1 based on operation and maintenance operation 1, and determines, based on semantic label 1, the operation and maintenance tools that can implement operation and maintenance operation 1, including operation and maintenance tool 1 to operation and maintenance tool 4. The computing device uses the second model to determine, based on result 1, that operation and maintenance tool 2 can implement operation and maintenance operation 1 and obtain operation and maintenance result 1. The computing device uses the second model to determine the calling specification 2 corresponding to operation and maintenance tool 2. Calling specification 2 includes formal parameter 2 and actual parameter 2. Based on the output of the first model and the second model, the computing device determines that the operation and maintenance operation that implements the operation and maintenance intent description information includes operation and maintenance operation 1, the operation and maintenance tool that executes operation and maintenance operation 1 is operation and maintenance tool 2, and the calling specification that calls operation and maintenance tool 2 is calling specification 2.

[0143] The above example illustrates the process by which a computing device obtains a target operation and maintenance (O&M) process using the first and second models. In some possible examples, the computing device may also use fewer or more models to obtain the target O&M process; this application does not limit this. When the computing device uses fewer models (such as the third model) to obtain the target O&M process, the computing device inputs a prompt to the third model, which then directly obtains the target O&M process based on the prompt. When the computing device uses more models to obtain the target O&M process, the computing device can use more models to implement the functions implemented by the first or second model. The specific functions implemented by the more models can be adjusted according to the needs of the actual application; this application does not limit this. In some possible examples, when the computing device uses two models of the same type but different parameter quantities (such as the first model and the second model) to determine the target O&M process, this can also be referred to as using a cerebellum-brain architecture to determine the target O&M process.

[0144] In some possible scenarios, the computing device can also provide prompts. These prompts indicate the execution status of the operational intent description, such as: the target operational process corresponding to the user-input operational intent description has been determined. In this way, the user can determine the progress of the computing device in determining the target operational process based on the operational intent description, improving the user-friendliness of the interaction.

[0145] In some possible scenarios, after determining the target operation and maintenance process corresponding to the operation and maintenance intent description information, the computing device may also display task completion information. The task completion information indicates that the operation and maintenance intent description information has been implemented, such as the task being terminated via `exit`.

[0146] In some possible scenarios, the presentation form of the prompt information includes, but is not limited to, text, picture, audio, video, and the like. The computing device can provide the prompt information in various presentation manners, including but not limited to, displaying a prompt interface, providing prompt audio, and the like.

[0147] In some possible scenarios, the computing device can further provide a first interface. The first interface includes at least one of the following: the user inputted operation and maintenance intention description information, the prompt word generated according to the operation and maintenance intention description information, the operation and maintenance tool description information generated by the first model to implement the operation and maintenance intention description information, the operation and maintenance operation description information generated by the first model to implement the operation and maintenance intention, the result description information generated by the first model to implement the operation and maintenance intention, or the calling manner of the target operation and maintenance tool, and the like.

[0148] The above describes the operation and maintenance process determination method provided by the present application. The following describes the method for constructing the operation and maintenance knowledge base provided by the present application.

[0149] In a scenario requiring operation and maintenance, an operation and maintenance personnel disassembles an operation and maintenance intention into at least one operation and maintenance operation, and calls an operation and maintenance tool to implement the operation and maintenance intention. To ensure that the operation and maintenance time does not exceed the operation and maintenance time limit, the operation and maintenance personnel often continuously calls the operation and maintenance tool to implement the at least one operation and maintenance operation. Based on this, the computing device can acquire, according to an operation record recording the operation and maintenance personnel implementing the operation and maintenance intention, the operation and maintenance intention, the at least one operation and maintenance operation and the operation sequence for implementing the operation and maintenance intention, and the operation and maintenance tool and the calling manner of the operation and maintenance tool for implementing the at least one operation and maintenance operation, and construct the operation and maintenance knowledge base based on the acquired information. The process of the computing device constructing the operation and maintenance knowledge base is described below in the related description of Figure 5 .

[0150] Figure 5 A flowchart of a method for constructing an operation and maintenance knowledge base provided by the present application is shown in Figure 5 , which includes the following (A) to (C).

[0151] (A) The computing device acquires at least one operation and maintenance record.

[0152] The operation and maintenance record is used to indicate that, in a historical time period, an operation and maintenance personnel calls an operation and maintenance tool to implement an operation and maintenance operation according to an operation and maintenance intention. In some possible scenarios, the operation and maintenance record can also be referred to as a historical operation record.

[0153] In this case, the computing device obtains the historical operation record, and determines at least one second operation according to the historical operation record. Specifically, the computing device determines the execution time of the operation corresponding to the historical operation record. The computing device selects the operation whose execution time is within the sliding time window. The computing device determines the operations within the specified number of sliding time windows as continuous operations, and determines the above continuous operations as the at least one second operation. That is, the at least one second operation can be considered as the operations performed to achieve an operation intention. The time length included in the sliding time window and the specified number of sliding time windows can be preset or set according to actual application requirements, which is not limited in the present application.

[0154] (B) The computing device generates an operation flow corresponding to the at least one operation record according to the at least one operation record.

[0155] The operation flow corresponding to the at least one operation record includes: an operation intention corresponding to the at least one operation record, at least one second operation corresponding to the at least one operation record and an operation sequence, at least one second operation tool for implementing the at least one second operation, and a calling mode of the at least one second operation tool.

[0156] The computing device can generate the operation flow by using the following (b1) to (b3).

[0157] (b1) The computing device determines the tool dependency relationship between the at least one second operation tool according to the operation dependency relationship between the at least one second operation corresponding to the at least one operation record.

[0158] The computing device determines the execution time of the at least one second operation corresponding to the at least one operation record. The computing device determines the operation dependency relationship between the at least one second operation according to the execution time of the at least one second operation. The computing device determines the second operation tool for implementing each second operation in the at least one second operation. The computing device determines the tool dependency relationship between the at least one second operation tool according to the operation dependency relationship between the at least one second operation.

[0159] Exemplarily, the at least one second operation corresponding to the at least one operation record is: operation 1 to operation 6. The execution time corresponding to operation 1 to operation 6 is: t1 to t6. The later time point in t1 to t6 is later than the former time point. The operation tool for implementing operation 1 to operation 6 is: tool 1 to tool 6. In this case, the computing device can determine that the operation dependency relationship corresponding to operation 1 to operation 6 is: operation 1-operation 2-operation 3-operation 4-operation 5-operation 6. And the computing device determines the tool dependency relationship corresponding to tool 1 to tool 6 according to the operation dependency relationship corresponding to operation 1 to operation 6. The tool dependency relationship corresponding to tool 1 to tool 6 is: tool 1-tool 2-tool 3-tool 4-tool 5-tool 6.

[0160] (b2) The computing device determines the operation intent corresponding to the at least one operation record according to the at least one second operation and the tool dependency relationship.

[0161] The computing device determines the dependency type of the tool dependency relationship. Wherein the dependency type includes one of: linear, parallel, tree, directed acyclic graph or combination. And the computing device inputs the at least one second operation and the tool dependency relationship into the first model according to the prompt word template corresponding to the dependency type, to obtain the operation intent corresponding to the at least one operation record.

[0162] (b3) The computing device determines the operation flow corresponding to the at least one operation record according to the operation intent corresponding to the at least one operation record, the at least one second operation, the operation dependency relationship, the at least one second operation tool, and the tool dependency relationship.

[0163] The computing device generates an operation triple according to the operation intent corresponding to the at least one operation record, the at least one second operation, the operation dependency relationship, the at least one second operation tool, and the tool dependency relationship. The operation triple is: <the operation intent corresponding to the at least one operation record, the at least one second operation for implementing the operation intent and the operation order, the at least one second operation tool for implementing the at least one second operation and the calling manner of the operation tool>. The computing device determines the operation triple as the operation triple corresponding to the operation flow.

[0164] (C) The computing device constructs an operation knowledge base according to the operation flow corresponding to the at least one operation record.

[0165] The computing device adds the operation flow corresponding to the at least one operation record to the operation knowledge base.

[0166] In some possible scenarios, the computing device can also update the operations and maintenance (O&M) knowledge base online. Specifically, if the first and second information do not match, the first information of the alternative O&M process instruction is replaced with the second information to obtain the updated O&M knowledge base. The first information includes: the O&M operations performed to implement the O&M intent corresponding to the O&M intent description information and the method of invoking O&M tools. The second information includes: the O&M operations performed to implement the O&M intent description information and the method of invoking O&M tools.

[0167] For example, the alternative operation and maintenance process determined based on the operation and maintenance intent description information is operation and maintenance process 1. The operation and maintenance triplet corresponding to operation and maintenance process 1 is operation and maintenance three-dimensional group 1. Operation and maintenance three-dimensional group 1 is <operation and maintenance intent>. Figure 1 The operation and maintenance (O&M) operations described in the O&M knowledge base are: O&M Operation 1 - O&M Operation 2 - O&M Operation 3 - O&M Operation 4, O&M Tool 1 - O&M Tool 2 - O&M Tool 3 - O&M Tool 4. The O&M operations performed to implement the O&M intent description information are: O&M Operation 1 - O&M Operation 2 - O&M Operation 3 - O&M Operation 5. The O&M tool invocation method to implement the O&M intent description information is: O&M Tool 1 - O&M Tool 2 - O&M Tool 3 - O&M Tool 5. In this case, the O&M triplet 1 in the O&M knowledge base is changed from <O&M Intent>. Figure 1 Operation and maintenance operation 1-Operation and maintenance operation 2-Operation and maintenance operation 3-Operation and maintenance operation 4, Operation and maintenance tool 1-Operation and maintenance tool 2-Operation and maintenance tool 3-Operation and maintenance tool 4> updated to <Operation and maintenance intention> Figure 1 Operation and maintenance 1-Operation and maintenance 2-Operation and maintenance 3-Operation and maintenance 5, Operation and maintenance tools 1-Operation and maintenance tools 2-Operation and maintenance tools 3-Operation and maintenance tools 5>.

[0168] The above text combined Figure 5 The construction process of the operation and maintenance knowledge base provided in this application is described below, in conjunction with... Figure 6 This application describes the process of constructing constraint decoding rules.

[0169] Figure 6 This application provides a flowchart illustrating the process of constructing constraint decoding rules, as shown below. Figure 6 As shown, the process includes the following (1) to (6).

[0170] (1) The computing device obtains detailed information about multiple operation and maintenance tools.

[0171] The detailed information of the operation and maintenance tool includes but is not limited to a tool identifier, a parameter list, a parameter constraint, a use scenario, and a tool description. The tool identifier is used to indicate the operation and maintenance tool, including but not limited to a tool name, a tool code, a tool abbreviation, and the like. The parameter list includes but is not limited to a formal parameter list and an actual parameter list. The parameter constraint is used to indicate rules that should be met by a parameter type, a format, a range, a number, a dependency relationship, and the like of a parameter (such as a formal parameter and an actual parameter) received by the operation and maintenance tool. The use scenario is used to indicate an application range of the operation and maintenance tool, including but not limited to a function supported by the operation and maintenance tool, an operation and maintenance intention supported by the operation and maintenance tool, a fault scenario to which the operation and maintenance tool is applied, and the like. The tool description includes but is not limited to a function, a use, a characteristic, a use scenario, a core capability, and key information of the operation and maintenance tool.

[0172] (2) The computing device initializes semantic labels of the operation and maintenance tools according to the detailed information of the plurality of operation and maintenance tools.

[0173] The computing device can initialize semantic labels of the plurality of operation and maintenance tools in a label determination manner according to the detailed information of the plurality of operation and maintenance tools. For related descriptions of the label determination manner, please refer to the foregoing description, which is not repeated here. One operation and maintenance tool can correspond to one or more semantic labels. For example, the plurality of operation and maintenance tools include operation and maintenance tool 1 to operation and maintenance tool n. In this case, the computing device determines the semantic labels corresponding to each of the operation and maintenance tool 1 to operation and maintenance tool n.

[0174] (3) The computing device establishes tool label constraint rules based on the semantic labels corresponding to each of the plurality of operation and maintenance tools.

[0175] The tool label constraint rules include the plurality of operation and maintenance tools and the semantic labels corresponding to each of the plurality of operation and maintenance tools. One operation and maintenance tool corresponds to one or more semantic labels, and one semantic label corresponds to one or more operation and maintenance tools.

[0176] (4) The computing device establishes tool formal parameter constraint rules according to the detailed information of the plurality of operation and maintenance tools.

[0177] The tool formal parameter constraint rules include the operation and maintenance tools and the formal parameters corresponding to the operation and maintenance tools.

[0178] (5) The computing device establishes actual parameter constraint rules according to the detailed information of the plurality of operation and maintenance tools.

[0179] The tool formal parameter actual parameter constraint rules include the formal parameters of the operation and maintenance tools and the actual parameters corresponding to the formal parameters of the operation and maintenance tools.

[0180] (6) The computing device constructs constraint decoding rules based on the tool label constraint rules, the tool formal parameter constraint rules, and the actual parameter constraint rules.

[0181] The constraint decoding rule is a rule that the model complies with when determining the next token according to the current token.

[0182] In some possible scenarios, the computing device can also update the constraint decoding rule. For example, the computing device updates the constraint decoding rule after generating the target operation and maintenance process. Specifically, the computing device updates the calling manner of the target operation and maintenance tool to the calling manner of the target operation and maintenance tool generated based on the constraint decoding rule. The updating of the constraint decoding rule by the computing device can refer to updating at least one of the tool tag constraint rule, the tool parameter constraint rule, or the parameter argument constraint rule of the constraint decoding rule.

[0183] For example, the computing device adjusts the semantic tag corresponding to the target operation and maintenance tool in the constraint decoding rule to an updated semantic tag. The updated semantic tag can be determined according to at least one of the operation and maintenance intention description information, the at least one first operation and maintenance operation for implementing the operation and maintenance intention description information, or the at least one first operation and maintenance tool for implementing the at least one first operation and maintenance operation. For example, the semantic tag corresponding to the target operation and maintenance tool in the constraint decoding rule is semantic tag 1, and the operation and maintenance intention indicated by the semantic tag 1 is operation and maintenance intention. Figure 1 The computing device determines that the target operation and maintenance tool is used to implement the operation and maintenance intention indicated by the operation and maintenance intention description information. Figure 2 In this case, the computing device can update the semantic tag of the target operation and maintenance tool in the constraint decoding rule from semantic tag 1 to semantic tag 2. The operation and maintenance intention indicated by the semantic tag 2 is operation and maintenance intention.

[0184] For another example, the computing device adjusts the parameter corresponding to the target operation and maintenance tool in the constraint decoding rule to an updated parameter. The updated parameter can be determined according to at least one of the operation and maintenance intention description information, the at least one first operation and maintenance operation for implementing the operation and maintenance intention description information, or the at least one first operation and maintenance tool for implementing the at least one first operation and maintenance operation. For example, the parameter corresponding to the target operation and maintenance tool in the constraint decoding rule is parameter 1. The target operation and maintenance process determined by the computing device indicates that the parameter of the target operation and maintenance tool is parameter 2. In this case, the computing device updates the parameter 1 of the target operation and maintenance tool in the constraint decoding rule to the parameter 2.

[0185] For another example, the computing device adjusts the parameter corresponding to the target operation and maintenance tool in the constraint decoding rule to an updated parameter. The updated parameter can be determined according to at least one of the operation and maintenance intention description information, the at least one first operation and maintenance operation for implementing the operation and maintenance intention description information, or the at least one first operation and maintenance tool for implementing the at least one first operation and maintenance operation. For example, the parameter corresponding to the target operation and maintenance tool in the constraint decoding rule is parameter 1. The target operation and maintenance process determined by the computing device indicates that the parameter of the target operation and maintenance tool is parameter 2. In this case, the computing device updates the parameter 1 of the target operation and maintenance tool in the constraint decoding rule to the parameter 2.

[0186] The above text combined Figure 4 to Figure 6 The method provided in this application has been described. The following is a brief description of the method provided in this application, taking the model including the first model and the second model as an example.

[0187] Figure 7 A schematic block diagram illustrating a method for determining an operation and maintenance process provided in this application, such as... Figure 7 As shown, the computing device can execute the following (①) to (③) steps to determine the target operation and maintenance process that implements the operation and maintenance intent description.

[0188] (①) The computing device constructs constraint decoding rules.

[0189] Please continue reading Figure 7 ,like Figure 7 As shown, the computing device can perform the following (11) to (17) to construct constraint decoding rules.

[0190] (11) Computing device acquisition tool library.

[0191] The tool library includes detailed information on multiple operation and maintenance tools. For detailed descriptions of these tools, please refer to the above text, which will not be repeated here.

[0192] (12) The computing device obtains detailed information on multiple operation and maintenance tools from the tool library.

[0193] (13) The computing device determines the detailed information of multiple operation and maintenance tools and initializes the semantic tags of the operation and maintenance tools.

[0194] (14) The computing device establishes tool tag constraint rules based on the semantic tags corresponding to each of the multiple operation and maintenance tools.

[0195] (15) The computing device establishes tool parameter constraint rules based on the detailed information of multiple operation and maintenance tools.

[0196] (16) The computing device establishes formal and actual parameter constraint rules based on the detailed information of multiple operation and maintenance tools.

[0197] (17) The computing device constructs constraint decoding rules based on tool label constraint rules, tool formal parameter constraint rules and formal parameter actual parameter constraint rules.

[0198] (②) Build an operation and maintenance knowledge base for computing devices.

[0199] Please continue reading Figure 7 ,like Figure 7 As shown, the computing device can perform the following (21) to (26) steps to build an operations and maintenance knowledge base.

[0200] (21) The computing device acquires historical operation records.

[0201] (22) The computing device obtains at least one operation record according to historical operation records.

[0202] The computing device determines at least one historical operation record by using a specified number of sliding time windows, and determines the at least one historical operation record as at least one operation record. The computing device determines at least one operation record corresponding to the specified number of sliding time windows as operation records corresponding to one operation intent.

[0203] (23) The computing device obtains an operation dependency relationship between at least one second operation corresponding to the at least one operation record.

[0204] (24) The computing device determines a tool dependency relationship between at least one second tool for implementing at least one second operation according to the operation dependency relationship.

[0205] (25) The computing device determines a dependency type of the tool dependency relationship.

[0206] (26) The computing device determines an operation intent corresponding to the at least one operation record according to the at least one second operation, the tool dependency relationship, and the dependency type of the tool dependency relationship.

[0207] (27) The computing device determines an operation process corresponding to the at least one operation record according to the operation intent corresponding to the at least one operation record, the at least one second operation, the operation dependency relationship, the at least one second tool for implementing the at least one second operation, and the tool dependency relationship.

[0208] (28) The computing device constructs an operation knowledge base according to the operation process corresponding to the at least one operation record.

[0209] For more description of the computing device constructing the operation knowledge base, please refer to the description of (28) above, which will not be repeated here. Figure 5

[0210] According to actual application needs, the computing device can execute (①) first and then execute (②), or the computing device can execute (①) and (②) simultaneously, and the order of the computing device executing (①) and (②) is not limited in the present application.

[0211] (③) The computing device obtains a target operation process for implementing the operation intent description information input by the user according to the constraint decoding rule and the operation knowledge base.

[0212] Please continue to refer to Figure 7 , as shown in Figure 7 , the computing device can execute the following (31) to (34) to obtain the target operation process. ​

[0213] (31) The computing device obtains the user inputted operation and maintenance intention description information.

[0214] (32) The computing device generates a prompt word based on the operation and maintenance knowledge base and the operation and maintenance intention description information.

[0215] (33) The computing device inputs the prompt word into a first model to generate operation and maintenance tool description information, operation and maintenance operation description information, and result description information for implementing the operation and maintenance intention description information.

[0216] The operation and maintenance tool description information, operation and maintenance operation description information, and result description information for implementing the operation and maintenance intention description information include multiple steps. One step includes an operation and maintenance operation description information, operation and maintenance tool description information for implementing an operation and maintenance operation, and result description information for implementing the operation and maintenance operation by calling the operation and maintenance tool description information.

[0217] (34) The computing device inputs the first step for implementing the operation and maintenance intention description information into a second model to generate a calling mode of a target operation and maintenance tool corresponding to the first step.

[0218] After performing (34), the computing device can further perform (35) to (38) for implementing the first step according to the target operation and maintenance process.

[0219] (35) The computing device calls the target operation and maintenance tool according to the calling mode of the target operation and maintenance tool corresponding to the first step to obtain a result.

[0220] (36) The computing device analyzes the result to determine whether it is reasonable or unreasonable to implement the operation and maintenance intention description information by the first step.

[0221] (37) In the case of unreasonableness, the computing device adjusts the prompt word and re-executes (33) to (36) above.

[0222] (38) In the case of reasonableness, the computing device continues to perform a second step for implementing the operation and maintenance intention description information.

[0223] The computing device can perform the second step in the same way as performing the first step. The above operations are repeated until the operation and maintenance intention description information is implemented.

[0224] In some possible cases, the computing device can also update the operation and maintenance knowledge base, such as updating the operation and maintenance knowledge base based on the target operation and maintenance process. For more description of the computing device updating the operation and maintenance knowledge base, please refer to the above description, which is not repeated here.

[0225] In some possible cases, the computing device can also update the constraint decoding rule, such as updating the constraint decoding rule based on the target operation and maintenance process.

[0226] Figure 8 This application provides a schematic diagram of an updated constraint decoding rule, as shown below. Figure 8 As shown in (a), the computing device can perform the following (41) to (45) update constraint decoding rules.

[0227] (41) The target operation and maintenance process for computing devices to obtain and implement operation and maintenance intention description information.

[0228] (42) The computing device obtains the call results of the target operation and maintenance tool using the call specification.

[0229] (43) Evaluation call results of computing devices.

[0230] (44) If the call result meets the requirements, the computing device determines the updated semantic tag of the target operation and maintenance tool according to the target operation and maintenance process that implements the operation and maintenance intent description information.

[0231] (45) The computing device updates the semantic label corresponding to the target operation and maintenance tool in the constraint decoding rule to the updated semantic label.

[0232] For a more detailed description of how computing devices build and update constraint decoding rules, please see the section above. Figure 6 The relevant descriptions will not be repeated here.

[0233] The process of building and updating constraint decoding rules for computing devices has been described above. The process of computing devices implementing operation and maintenance intent description information based on the operation and maintenance knowledge base and the tool tag constraint rules in the constraint decoding rules is described below.

[0234] Please continue reading Figure 8 ,like Figure 8 As shown in (b), the computing device acquires an operations and maintenance (O&M) toolset containing detailed information on multiple O&M tools, and constructs tool tag constraint rules, tool formal parameter constraint rules, and formal and actual parameter constraint rules based on the O&M toolset. The computing device acquires an O&M knowledge base. Based on the tool tag constraint rules and the O&M knowledge base, the computing device determines third information. The third information includes: O&M tool description information, O&M operation description information, and result description information that realize the O&M intent description. The computing device inputs the O&M tool description information, O&M operation description information, and result description information that realize the O&M intent description into the second model to generate the calling method of the target O&M tool. The second model is used to generate the calling specification of the O&M tool based on: the tool tag constraint rules, tool formal parameter constraint rules, and formal and actual parameter constraint rules, and the O&M tool description information, O&M operation description information, and result description information that realize the O&M intent description.

[0235] The above describes the method for determining the operation and maintenance process provided in this application. The following section uses specific examples to illustrate the method provided in this application.

[0236] Figure 9 A flowchart of an example of a method for determining an operation and maintenance process provided by the present application is shown in FIG. 1(a). After a user inputs operation and maintenance intention description information “Help me analyze what commonalities and abnormalities these hosts have, and the host information is as follows: Internet Protocol address ip 1 of host 1 to ip n of host n”, the computing device provides a first interface. The first interface includes at least one of the following: the user-input operation and maintenance intention description information, a prompt word, operation and maintenance tool description information for implementing the operation and maintenance intention description information, operation and maintenance operation description information generated by a first model for implementing the operation and maintenance intention, result description information generated by the first model for implementing the operation and maintenance intention, or a calling manner for a target operation and maintenance tool. Figure 9

[0237] Specifically, the user-input operation and maintenance intention description information is “User input: Help me analyze what commonalities and abnormalities these hosts have, and the host information is as follows: Internet Protocol address ip 1 of host 1 to ip n of host n”.

[0238] The prompt word is “Prompt word prompt: current operation and maintenance tool set {tool 1 to tool n}; alternative operation and maintenance processes <batch host commonalities and abnormality analysis, query xx host location-query xx host status-query xx host packet loss-query xx associated switch-query xx switch alarm-query xx switch change, operation and maintenance tool 6-operation and maintenance tool 7-operation and maintenance tool 9-operation and maintenance tool 3-operation and maintenance tool 1-operation and maintenance tool 2>”, and based on the above information, complete the user request: analyze what commonalities and abnormalities the following hosts have, and the host information is as follows: Internet Protocol address ip 1 of host 1 to ip n of host n”.

[0239] The operation and maintenance tool description information for implementing the operation and maintenance intention description information, the operation and maintenance operation description information generated by the first model for implementing the operation and maintenance intention, and the result description information generated by the first model for implementing the operation and maintenance intention include: first step to sixth step. The first step to the sixth step are displayed in the first interface as follows.

[0240] For the first step, the computing device displays the following information according to the content generated by the first model: “First step: according to the user's question, the host location should be checked, and the host location query tool should be used to check the host location. The query results should be analyzed from the perspective of the geographical location of the host, that is, the first step should check the host information, and according to the user-given ip 1 to ip n, call operation and maintenance tool 6 to query the physical location information of the host corresponding to ip 1 to ip n”.

[0241] ​For the first step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 6, formal parameter: {formal parameter 11, … formal parameter 1m}, actual parameter: {actual parameter 11, … actual parameter 1m}}".

[0242] For the first step, the computing device executes the call of the target operation and maintenance tool according to the call mode of the target operation and maintenance tool obtained by the second model to obtain a call result, and the computing device displays the following information according to the call result: "Result: {ip: ip 1 to ip n, address: location 1 to location n}".

[0243] For the second step, the computing device displays the following information according to the content generated by the first model and the execution result of the second step: "The second step should call operation and maintenance tool 7 to query the cloud agent state of the host to be in-band or out-of-band state".

[0244] For the second step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 7, formal parameter: {formal parameter 21, … formal parameter 2n}, actual parameter: {actual parameter 21, … actual parameter 2n}}".

[0245] For the second step, the computing device executes the call of the target operation and maintenance tool according to the call mode of the target operation and maintenance tool obtained by the second model to obtain a call result, and the computing device displays the following information according to the call result: "Result: {ip: ip 1 to ip n, state: state 1 to state n}".

[0246] For the third step, the computing device displays the following information according to the content generated by the first model and the execution result of the third step: "The third step should call operation and maintenance tool 9 to query the packet loss of the host".

[0247] For the third step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 9, formal parameter: {formal parameter 31, … formal parameter 3i}, actual parameter: {actual parameter 31, … actual parameter 3i}}".

[0248] For the third step, the computing device executes the call of the target operation and maintenance tool according to the call mode of the target operation and maintenance tool obtained by the second model to obtain a call result, and the computing device displays the following information according to the call result: "Result: {ip: ip 1 to ip n, state: packet loss information 1 to packet loss information n}".

[0249] For the fourth step, the computing device displays the following information according to the content generated by the first model and the execution result of the fourth step: "The fourth step should call the operation and maintenance tool 9 to query the associated switch".

[0250] For the fourth step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 9, parameter: {parameter 41, … parameter 4j}, actual parameter: {actual parameter 41, … actual parameter 4j}}".

[0251] For the fourth step, the computing device obtains the call mode of the target operation and maintenance tool according to the second model, executes the call of the target operation and maintenance tool to obtain the call result, and the computing device displays the following information according to the call result: "Result: {ip: ip 1 to ip n, state: associated switch information 1 to associated switch information n}".

[0252] For the fifth step, the computing device displays the following information according to the content generated by the first model and the execution result of the fifth step: "The fifth step should call the operation and maintenance tool 1 to query the switch alarm".

[0253] For the fifth step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 1, parameter: {parameter 51, … parameter 5k}, actual parameter: {actual parameter 51, … actual parameter 5k}}".

[0254] For the fifth step, the computing device obtains the call mode of the target operation and maintenance tool according to the second model, executes the call of the target operation and maintenance tool to obtain the call result, and the computing device displays the following information according to the call result: "Result: {ip: ip 1 to ip n, state: switch alarm information 1 to switch alarm information n}".

[0255] For the sixth step, the computing device displays the following information according to the content generated by the first model and the execution result of the sixth step: "The sixth step should call the operation and maintenance tool 2 to query the switch change".

[0256] For the sixth step, the computing device displays the following information according to the content generated by the second model: "Call mode: {semantic label label: label 1, tool identifier: operation and maintenance tool 2, parameter: {parameter 61, … parameter 6y}, actual parameter: {actual parameter 61, … actual parameter 6y}}".

[0257] For the sixth step, the computing device performs the calling of the target operation and maintenance tool according to the calling mode of the target operation and maintenance tool obtained according to the second model, obtains a calling result, and displays the following information according to the calling result: "Result: {ip: ip 1 to ip n, status: switch change 1 to switch change n}".

[0258] In some possible cases, the computing device can also display prompt information as shown in (a) of FIG. 6, for example, display "Query result: confirmation that the query has been completed, based on the above information, the comprehensive analysis result is that host 1 to host n are all deployed in location 1, host 1 to host n are all in the in-band state, host 1 to host n are all not packet loss, host 1 to host n are associated with switch 1, host 1 to host n are not associated with switch alarm, and host 1 to host n are not associated with switch change". Figure 9 In some possible cases, the computing device can also display task end information as shown in (a) of FIG. 6, for example, display "Task completion termination exit".

[0259] Figure 9 In some possible cases, the computing device can also display task end information as shown in (a) of FIG. 6, for example, display "Task completion termination exit".

[0260] The above describes the first interface provided by the computing device in combination with (a) of FIG. 6, and the following describes the process performed by the computing device after the computing device obtains the operation and maintenance intention description information input by the user in combination with (b) of FIG. 6. Figure 9 Figure 9 The above describes the first interface provided by the computing device in combination with (a) of FIG. 6, and the following describes the process performed by the computing device after the computing device obtains the operation and maintenance intention description information input by the user in combination with (b) of FIG. 6.

[0261] In this example, the target operation and maintenance tools are operation and maintenance tool 1, operation and maintenance tool 2, operation and maintenance tool 3, operation and maintenance tool 8, and operation and maintenance tool 7, and the example is described in combination with (a) of FIG. 6. Figure 9 Figure 9 ​​​As shown in (b), the computing device constructs decoding constraint rules including those shown in Table 1 above. The computing device constructs an operations and maintenance (O&M) knowledge base. After obtaining the O&M intent description information input by the user, the computing device, based on the O&M knowledge base and decoding constraint rules, uses the first model to generate O&M tool description information, O&M operation description information, and result description information that implement the O&M intent description information. The computing device inputs the aforementioned O&M tool description information, O&M operation description information, and result description information that implement the O&M intent description information into the second model to generate a calling method for the target O&M tool. The calling method for the target O&M tool includes, but is not limited to: the tool identifier of the target O&M tool, the formal parameters of the target O&M tool, and the actual parameters corresponding to the formal parameters. For example, the tool identifiers of the target O&M tool include: O&M tool 1, O&M tool 2, O&M tool 3, O&M tool 8, and O&M tool 7. The formal parameters of O&M tool 1 include: region, cloud service, alarm level, and time. The actual parameters corresponding to the formal parameters of O&M tool 1 include: region 1, datacenter network (dcn), emergency, and within one hour. The parameters for Operations and Maintenance Tool 2 include: Change Type, Change Level, Change Quantity, and Change Order Number. The corresponding actual parameters for Operations and Maintenance Tool 2 include: Regular Change, Level 1, 50, and Order Number 1. The parameters for Operations and Maintenance Tool 3 include: Device Type, IP Address, Required Keywords, and Region. The corresponding actual parameters for Operations and Maintenance Tool 3 include: Switch, IP1, Topology, and Region 2. The parameters for Operations and Maintenance Tool 8 include: Region, Availability Zone (AZ), Container Pod, and Monitoring Metrics. The corresponding actual parameters for Operations and Maintenance Tool 8 include: Region 3, Availability Zone 1, Container Pod 1, and Packet Loss Rate. The parameters for Operations and Maintenance Tool 7 include: Device IP, Device Role, Interface Name, and Switch Port Status. The corresponding actual parameters for Operations and Maintenance Tool 7 include: IP1, Access Switch, Name 1, and Normal Operation.

[0262] The above text combined Figure 9 The method provided in this application has been illustrated by example, and the following examples illustrate the scenarios in which the method provided in this application can be applied.

[0263] The method provided in this application can be applied to various scenarios, including but not limited to: system risk control scenarios, rapid fault recovery scenarios, etc. In some possible examples, system risk control scenarios may also be called risk control scenarios, and rapid fault recovery scenarios may also be called fast recovery scenarios, etc. The following will combine... Figure 10 The following are exemplary descriptions of the scenarios in which the methods provided in this application can be applied.

[0264] Figure 10 Example diagrams illustrating the application scenarios of the operation and maintenance process methods provided in this application, such as... Figure 10As shown in (a), the risk control scenario can include: topology generation, data coloring, and data analysis. The computing device can generate the required topology links based on the inspection results and risk investigation results. The computing device performs data coloring on the aforementioned operational data to obtain coloring results. The computing device generates analysis results and analysis details based on the coloring results. Furthermore, the computing device can also generate analysis components based on the analysis results and push analysis reports.

[0265] In the above process, the computing device can use the operation and maintenance process determination method provided in this application to generate topology links based on inspection results and troubleshooting results, perform data coloring, generate analysis results and analysis details, analysis reports, etc.

[0266] Please continue reading Figure 10 ,like Figure 10 As shown in (b), the rapid recovery scenario can include: determining the rapid recovery execution entity / rapid recovery clue organization, fault investigation / business escape, and recovery handling / retrospective analysis. Specifically, the computing device performs the following process to organize rapid recovery clues: the computing device can respond to user commands or / or automatically trigger the acquisition of multiple instances of fault management, and analyze the commonalities of multiple instances to obtain a first analysis result. Based on the first analysis result, the computing device analyzes whether dependent resources are abnormal or not, whether performance is at / not at a bottleneck, and whether the device's health status is healthy / sub-healthy / unhealthy, obtaining a second analysis result. The computing device performs the following process to implement the fault investigation / business escape process: specifically, the computing device identifies the fault scenario based on the first analysis result or the first and second analysis results. Based on the fault scenario, the computing device determines the fault content, loss-stopping object, and loss-stopping strategy. After the computing device determines the fault content, loss-stopping object, and loss-stopping strategy, it can generate customer statements. The computing device performs the following process to implement the recovery handling / retrospective analysis: specifically, the computing device determines the recovery decision based on the fault content, loss-stopping object, and loss-stopping strategy. The computing device generates a summary report based on the recovery decision and performs a review and correction based on the report. After obtaining the first and second analysis results, the computing device can perform a progress summary / information synchronization to synchronize the fault handling progress. Similarly, the computing device can also perform a progress summary / information synchronization after identifying the fault scenario. It can also perform a progress summary / information synchronization after generating customer specifications.

[0267] In the above process, the computing device can use the operation and maintenance process determination method provided in this application to obtain the first analysis result, the second analysis result, identify the fault scenario, determine the fault content, determine the loss prevention object, determine the loss prevention strategy, generate customer specifications, etc.

[0268] In the present application, the computing device determines the candidate operation and maintenance process associated with the operation and maintenance intention description information according to the operation and maintenance knowledge base containing operation and maintenance knowledge, and determines the prompt word according to the above-mentioned candidate operation and maintenance process, so that the prompt word contains the operation and maintenance knowledge associated with the operation and maintenance intention description information. In this way, the computing device can obtain the operation and maintenance knowledge associated with the operation and maintenance intention description information according to the prompt word, without training the initial model with training data to obtain the operation and maintenance knowledge, thereby reducing the cost.

[0269] It can be understood that, in order to realize the functions in the above-mentioned embodiments, the computing device includes the hardware structure and / or software module corresponding to each function. Those skilled in the art should easily realize that, in combination with the units and method steps of each example described in the embodiments disclosed in the present application, the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is realized in the form of hardware or computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0270] The operation and maintenance process determination method provided by the present embodiment is described in detail above. Figure 1 to Figure 10 The operation and maintenance process determination device provided by the present embodiment will be described below. Figure 11

[0271] Figure 11 The structure diagram of the operation and maintenance process determination device provided by the present application is shown in FIG. 11. Figure 11 The operation and maintenance process determination device 1100 includes an acquisition module 1110 and a processing module 1120. The operation and maintenance process determination device 1100 can be used to realize the functions of the computing device in the operation and maintenance process determination method.

[0272] When the operation and maintenance process determination device shown in FIG. 11 is used to realize the functions of the computing device in the operation and maintenance process determination method, the operation and maintenance process determination device 1100 includes an acquisition module 1110 and a processing module 1120. The operation and maintenance process determination device 1100 can be the computing device 210 shown in FIG. 2, or the chip shown in FIG. 3. The acquisition module 1110 can be used to realize the functions implemented in S410 in the above-mentioned method embodiments, and the processing module 1120 can be used to realize the functions implemented in S420 to S440 in the above-mentioned method embodiments. Figure 11 Figure 2 Figure 3

[0273] ​​​​The acquisition module 1110 is configured to acquire user input operation and maintenance intention description information. The processing module 1120 is configured to determine, based on an operation and maintenance knowledge base, a candidate operation and maintenance process associated with the operation and maintenance intention description information. The operation and maintenance knowledge base includes at least one operation and maintenance process. One operation and maintenance process is used to indicate operation and maintenance operations performed to achieve one operation and maintenance intention and a calling manner of an operation and maintenance tool. The candidate operation and maintenance process is used to indicate operation and maintenance operations performed to achieve an operation and maintenance intention corresponding to the operation and maintenance intention description information and a calling manner of an operation and maintenance tool. The processing module 1120 is further configured to generate a prompt word according to the candidate operation and maintenance process and meta-information of an operation and maintenance tool in an operation and maintenance tool set. The operation and maintenance tool in the operation and maintenance tool set is determined according to the operation and maintenance intention description information. The processing module 1120 is further configured to input the prompt word into a model to obtain a target operation and maintenance process. The target operation and maintenance process is used to indicate target operation and maintenance operations performed to achieve the operation and maintenance intention description information and a calling manner of a target operation and maintenance tool.

[0274] In some possible cases, the operation and maintenance knowledge base includes at least one operation and maintenance triple. One operation and maintenance triple includes an operation and maintenance intention, at least one first operation and maintenance operation performed to achieve the operation and maintenance intention and an operation sequence, and at least one first operation and maintenance tool performing the at least one first operation and maintenance operation and a calling manner of the at least one first operation and maintenance tool. One operation and maintenance triple corresponds to one operation and maintenance process.

[0275] In some possible cases, the target operation and maintenance process corresponds to a target operation and maintenance triple, and the at least one operation and maintenance triple in the operation and maintenance knowledge base includes or does not include the target operation and maintenance triple.

[0276] In some possible cases, the model includes a first model and a second model. The processing module 1120 is specifically configured to input the prompt word into the first model to generate operation and maintenance tool description information, operation and maintenance operation description information and result description information used to achieve the operation and maintenance intention description information. The processing module 1120 is further specifically configured to input the operation and maintenance tool description information, the operation and maintenance operation description information and the result description information into the second model to generate the calling manner of the target operation and maintenance tool. The second model is used to generate the calling manner of the target operation and maintenance tool according to a constraint decoding rule. The constraint decoding rule includes a plurality of operation and maintenance tools and at least one semantic label corresponding to the plurality of operation and maintenance tools. One semantic label corresponds to at least one of the plurality of operation and maintenance tools, one operation and maintenance tool corresponds to at least one of the at least one semantic label, and the semantic label is used to indicate an application range of the operation and maintenance tool.

[0277] In some possible implementations, the processing module 1120 is further configured to: obtain at least one operation and maintenance record. One operation and maintenance record includes: a second operation and maintenance operation, a tool identifier of a second operation and maintenance tool for implementing the second operation and maintenance operation, and a result of calling the second operation and maintenance tool to perform the second operation and maintenance operation. The processing module 1120 is further configured to: generate, according to the at least one operation and maintenance record, an operation and maintenance process corresponding to the at least one operation and maintenance record. The operation and maintenance process corresponding to the at least one operation and maintenance record includes: an operation and maintenance intent corresponding to the at least one operation and maintenance record, at least one second operation and maintenance operation corresponding to the at least one operation and maintenance record and an operation sequence, and at least one second operation and maintenance tool for implementing the at least one second operation and maintenance operation and a calling manner of the at least one second operation and maintenance tool. The processing module 1120 is further configured to: construct, according to the operation and maintenance process corresponding to the at least one operation and maintenance record, an operation and maintenance knowledge base.

[0278] In some possible implementations, the processing module 1120 is specifically configured to: determine, according to an operation dependency relationship between the at least one second operation and maintenance operation corresponding to the at least one operation and maintenance record, a tool dependency relationship between at least one second operation and maintenance tool for implementing the at least one second operation and maintenance operation. The processing module 1120 is further specifically configured to: determine, according to the at least one second operation and maintenance operation and the tool dependency relationship, the operation and maintenance intent corresponding to the at least one operation and maintenance record. The processing module 1120 is further specifically configured to: determine, according to the operation and maintenance intent corresponding to the at least one operation and maintenance record, the at least one second operation and maintenance operation, the at least one second operation and maintenance tool, and the tool dependency relationship, the operation and maintenance process corresponding to the at least one operation and maintenance record.

[0279] In some possible implementations, the model includes a first model. The first model is used to obtain an operation and maintenance intent according to a tool dependency relationship between an operation and maintenance operation and an operation and maintenance tool for implementing the operation and maintenance operation. The processing module 1120 is specifically configured to: determine a dependency type of the tool dependency relationship. The dependency type includes one of: linear, parallel, tree, directed acyclic graph, or combination. The processing module 1120 is further specifically configured to: input, according to a prompt word template corresponding to the dependency type, the at least one second operation and maintenance operation and the tool dependency relationship into the first model to obtain the operation and maintenance intent corresponding to the at least one operation and maintenance record.

[0280] In some possible implementations, the processing module 1120 is further configured to: in a case where the first information and the second information do not match, replace the first information indicated by the alternative operation and maintenance process with the second information to obtain an updated operation and maintenance knowledge base. The first information is: an operation and maintenance operation performed to implement an operation and maintenance intent corresponding to operation and maintenance intent description information and a calling manner of an operation and maintenance tool. The second information is: an operation and maintenance operation performed to implement the operation and maintenance intent description information and a calling manner of the operation and maintenance tool.

[0281] In some possible implementation, the processing module 1120 is further configured to determine semantic labels corresponding to the plurality of operation and maintenance tools according to application ranges of the plurality of operation and maintenance tools. The target operation and maintenance tool belongs to the plurality of operation and maintenance tools. The processing module 1120 is further configured to construct constraint decoding rules based on the plurality of operation and maintenance tools and the semantic labels corresponding to the plurality of operation and maintenance tools.

[0282] In some possible implementation, the processing module 1120 is further configured to update the calling mode of the target operation and maintenance tool to the calling mode of the target operation and maintenance tool generated based on the constraint decoding rules.

[0283] In some possible implementation, the first model is configured to generate the operation and maintenance process according to the prompt word, and obtain the operation and maintenance intent according to the operation and maintenance operation and a dependency relationship between operation and maintenance tools implementing the operation and maintenance operation.

[0284] For more information about the obtaining module 1110 and the processing module 1120, please refer to the description of the computing device in the operation and maintenance process determination method above, which will not be repeated here.

[0285] In the case where the operation and maintenance process determination apparatus 1100 corresponds to the steps performed by the computing device in the operation and maintenance process determination method described in the embodiments of the present application, the above and other operations and / or functions of each module in the operation and maintenance process determination apparatus 1100 are respectively used to implement the method processes performed by the computing device in the foregoing figures.

[0286] The operation and maintenance process determination apparatus 1100 described above can be implemented by software or hardware.

[0287] In the case where the operation and maintenance process determination apparatus 1100 is implemented by a software module, the software module can be provided in multiple ways. For example, the software module can be provided to users for use through a cloud service subscription mode, and users can choose different subscription levels according to needs; for another example, the software module can also provide enterprise-level customized services with professional domain customization, interface personalization, and expansion functions according to the needs of users or enterprises.

[0288] In addition, the operation and maintenance process determination apparatus 1100 provided by the present application can also be provided as a value-added service to users, which is not limited herein.

[0289] In the case where the operation and maintenance process determination apparatus is implemented by hardware, the operation and maintenance process determination apparatus can be a computing device, a chip, or a processor. For specific implementation of the computing device, please refer to the description of the computing device in the foregoing figures, and for specific implementation of the chip and the processor, please refer to the description of the chip and the processor in the foregoing figures, which will not be repeated here. Figure 2 Figure 3

[0290] ​​The obtaining module 1110 and the processing module 1120 can be implemented by software, or can be implemented by hardware, or can be implemented by a combination of software and hardware, and the present application does not limit the implementation manner. Next, the implementation manner of the processing module 1120 will be introduced. Similarly, the implementation manner of the obtaining module 1110 can refer to the implementation manner of the processing module 1120.

[0291] As an example of a software functional unit, the processing module 1120 can include code running on a computing instance. The computing instance can include at least one of a physical host (computing device), a virtual machine, and a container. Further, the computing instance can be one or more. For example, the processing module 1120 can include code running on multiple hosts / virtual machines / containers. It should be noted that the multiple hosts / virtual machines / containers for running the code can be distributed in the same region, or can be distributed in different regions. Further, the multiple hosts / virtual machines / containers for running the code can be distributed in the same availability zone (AZ), or can be distributed in different AZs, and each AZ includes one data center or multiple data centers with similar geographical locations. Generally, one region can include multiple AZs.

[0292] Similarly, the multiple hosts / virtual machines / containers for running the code can be distributed in the same virtual private cloud (VPC), or can be distributed in multiple VPCs. Generally, one VPC is set in one region, and a communication gateway needs to be set in each VPC for cross-region communication between two VPCs in the same region and between VPCs in different regions, and the interconnection between VPCs is realized through the communication gateway.

[0293] As an example of a hardware functional unit, the processing module 1120 can include at least one computing device, such as a server. Alternatively, the processing module 1120 can also be a device implemented by an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), etc. The PLD can be implemented by a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0294] In the case where the processing module 1120 includes at least two computing devices, the at least two computing devices included in the processing module 1120 can be distributed in the same region or in different regions. The at least two computing devices included in the processing module 1120 can be distributed in the same AZ or in different AZs. Likewise, the at least two computing devices included in the processing module 1120 can be distributed in the same VPC or in multiple VPCs. The multiple computing devices can be any combination of servers, ASICs, PLDs, CPLDs, FPGAs, and GALs.

[0295] It should be noted that in other embodiments, the acquisition module 1110 can be configured to perform any step of the operation and maintenance process determination method, and the processing module 1120 can be configured to perform any step of the operation and maintenance process determination method. The steps responsible for implementation by the acquisition module 1110 and the processing module 1120 can be specified as needed, and the acquisition module 1110 and the processing module 1120 respectively implement different steps of the operation and maintenance process determination method to realize all functions of the operation and maintenance process determination apparatus 1100.

[0296] The method steps in this embodiment can be implemented by hardware, or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in a random access memory (RAM), a flash memory, a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically EPROM (EEPROM), a register, a hard disk, a mobile hard disk, a CD-ROM, or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC. In addition, the ASIC can be located in a computing device. Of course, the processor and the storage medium can also exist as discrete components in a network device or a terminal device.

[0297] The embodiments of the present application also provide a computing device cluster. The computing device cluster includes at least one computing device. The computing device can be a device with data processing capability.

[0298] As shown in Figure 12 ,Figure 12 A structural diagram of a computing device cluster is provided in the present application, and the computing device cluster includes at least one computing device 210. The same instructions for performing the operation and maintenance process determination method can be stored in the memory 212 of one or more computing devices 210 in the computing device cluster.

[0299] In some possible implementation manners, partial instructions for performing the operation and maintenance process determination method can also be respectively stored in the memory 212 of one or more computing devices 210 in the computing device cluster. In other words, the combination of one or more computing devices 210 can collectively execute the instructions for performing the operation and maintenance process determination method.

[0300] It should be noted that the memories 212 in different computing devices 210 in the computing device cluster can store different instructions, respectively used to perform part of the functions of the computing device in the operation and maintenance process determination method. That is, the instructions stored in the memories 212 in different computing devices 210 can implement the functions of one or more of the obtaining module 1110 and the processing module 1120.

[0301] In some possible implementation manners, one or more computing devices in the computing device cluster can be connected through a network. The network can be a wide area network or a local area network, etc. Figure 13 A possible implementation manner is shown. As shown in Figure 13 Figure 13 A connection diagram between computing devices is provided in the present application, and two computing devices 210A and 210B are connected through a network. Specifically, the communication interface in each computing device is connected to the network. In this type of possible implementation manner, the instructions stored in the memory 212 in the computing device 210A can implement the functions implemented by the obtaining module 1110. At the same time, the instructions stored in the memory 212 in the computing device 210B can implement the functions implemented by the processing module 1120.

[0302] The embodiments of the present application further provide a computer program product containing instructions. The computer program product can be software or a program product containing instructions, capable of running on a computing device or stored in any available medium. When the computer program product runs on at least one computing device, the at least one computing device is caused to perform the operation and maintenance process determination method.

[0303] ​The embodiment of the present application further provides a computer readable storage medium. The computer readable storage medium can be any available medium or data storage device that can be accessed by a computing device, such as a data center containing one or more available media. The available medium can be a magnetic medium, such as a floppy diskette, a hard disk drive, a magnetic tape, an optical medium, such as a digital video disc (DVD), or a semiconductor medium, such as a solid state drive, or the like. The computer readable storage medium includes instructions that instruct the computing device to perform the operation and maintenance process determination method.

[0304] The present application further provides a chip. The chip includes an interface circuit and a control circuit. The interface circuit is configured to obtain the operation and maintenance intention description information, and the control circuit is configured to implement the function of the computing device in the operation and maintenance process determination method.

[0305] In the above embodiment, the implementation can be achieved by software, hardware, firmware or any combination thereof, in whole or in part. When implemented by software, the implementation can be achieved in the form of a computer program product, in whole or in part. The computer program product includes one or more computer programs or instructions. When the computer programs or instructions are loaded and executed on a computer, the processes or functions described in the embodiment of the present application are executed in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment or other programmable devices. The computer programs or instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer programs or instructions can be transferred from one website, computer, server or data center to another by wired or wireless means. The computer readable storage medium can be any available medium accessible by a computer or a data storage device, such as a server, data center, etc., containing one or more available media. The available medium can be a magnetic medium, such as a floppy disk, a hard disk, a magnetic tape, an optical medium, such as a digital video disc (DVD), or a semiconductor medium, such as a solid state drive (SSD).

[0306] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An operation and maintenance process determination method, characterized in that, The method is executed by a computing device on which a model is deployed, and the method comprises: obtaining user inputted operation and maintenance intention description information; determining, based on an operation and maintenance knowledge base, a candidate operation and maintenance process associated with the operation and maintenance intention description information; wherein the operation and maintenance knowledge base comprises at least one operation and maintenance process, and one operation and maintenance process is used to indicate operation and maintenance operations performed to realize one operation and maintenance intention and a calling manner of an operation and maintenance tool; the candidate operation and maintenance process is used to indicate operation and maintenance operations performed to realize an operation and maintenance intention corresponding to the operation and maintenance intention description information and a calling manner of an operation and maintenance tool; generating prompt words according to the candidate operation and maintenance process and meta information of operation and maintenance tools in an operation and maintenance tool set; the operation and maintenance tools in the operation and maintenance tool set are determined according to the operation and maintenance intention description information; inputting the prompt words into the model to obtain a target operation and maintenance process; the target operation and maintenance process is used to indicate target operation and maintenance operations performed to realize the operation and maintenance intention description information and a calling manner of a target operation and maintenance tool.

2. The method of claim 1, wherein, The operation and maintenance knowledge base comprises at least one operation and maintenance triple; one operation and maintenance triple comprises an operation and maintenance intention, at least one first operation and maintenance operation performed to realize the operation and maintenance intention and an operation sequence, and at least one first operation and maintenance tool performed to realize the at least one first operation and maintenance operation and a calling manner of the at least one first operation and maintenance tool; one operation and maintenance triple corresponds to one operation and maintenance process.

3. The method of claim 2, wherein, The target operation and maintenance process corresponds to a target operation and maintenance triple, and the at least one operation and maintenance triple in the operation and maintenance knowledge base comprises or does not comprise the target operation and maintenance triple.

4. The method of claim 3, wherein, The model comprises a first model and a second model. The inputting the prompt words into the model to obtain the target operation and maintenance process comprises: inputting the prompt words into the first model to generate operation and maintenance tool description information, operation and maintenance operation description information and result description information implemented to realize the operation and maintenance intention description information; inputting the operation and maintenance tool description information, the operation and maintenance operation description information and the result description information into the second model to generate the calling manner of the target operation and maintenance tool; wherein the second model is used to generate the calling manner of the target operation and maintenance tool according to a constraint decoding rule, and the constraint decoding rule comprises a plurality of operation and maintenance tools and at least one semantic label corresponding to the plurality of operation and maintenance tools; one semantic label corresponds to at least one of the plurality of operation and maintenance tools, one operation and maintenance tool corresponds to at least one of the at least one semantic label, and a semantic label is used to indicate an application range of an operation and maintenance tool.

5. The method of any one of claims 1-4, wherein, before the determining, based on an operation and maintenance knowledge base, a candidate operation and maintenance process associated with the operation and maintenance intention description information, the method further comprises: obtaining at least one operation and maintenance record; one operation and maintenance record comprises a second operation and maintenance operation, a tool identifier of a first operation and maintenance tool implemented to realize the second operation and maintenance operation, and a result of calling the first operation and maintenance tool to perform the second operation and maintenance operation. According to the at least one operation and maintenance record, an operation and maintenance process corresponding to the at least one operation and maintenance record is generated; the operation and maintenance process corresponding to the at least one operation and maintenance record comprises: an operation and maintenance intention corresponding to the at least one operation and maintenance record, at least one second operation and operation sequence corresponding to the at least one operation and maintenance record, and at least one second operation and maintenance tool for implementing the at least one second operation and maintenance tool and a calling mode of the at least one second operation and maintenance tool; According to the operation and maintenance process corresponding to the at least one operation and maintenance record, the operation and maintenance knowledge base is constructed.

6. The method of claim 5, wherein, According to the at least one operation and maintenance record, an operation and maintenance process corresponding to the at least one operation and maintenance record is generated; the operation and maintenance process corresponding to the at least one operation and maintenance record comprises: an operation and maintenance intention corresponding to the at least one operation and maintenance record, at least one second operation and operation sequence corresponding to the at least one operation and maintenance record, and at least one second operation and maintenance tool for implementing the at least one second operation and maintenance tool and a calling mode of the at least one second operation and maintenance tool; According to the operation and maintenance process corresponding to the at least one operation and maintenance record, the operation and maintenance knowledge base is constructed.

6. The method of claim 5, wherein, According to the at least one operation and maintenance record, an operation and maintenance process corresponding to the at least one operation and maintenance record is generated; the operation and maintenance process corresponding to the at least one operation and maintenance record comprises: an operation and maintenance intention corresponding to the at least one operation and maintenance record, at least one second operation and operation sequence corresponding to the at least one operation and maintenance record, and at least one second operation and maintenance tool for implementing the at least one second operation and maintenance tool and a calling mode of the at least one second operation and maintenance tool; 7. The method of claim 6, wherein, The model comprises a first model, and the first model is used to obtain an operation and maintenance intention according to a tool dependency relationship between an operation and maintenance operation and an operation and maintenance tool for implementing the operation and maintenance operation; According to the at least one second operation and maintenance operation and the tool dependency relationship, the operation and maintenance intention corresponding to the at least one operation and maintenance record is determined, comprising: A dependency type of the tool dependency relationship is determined; the dependency type comprises one of the following: linear, parallel, tree, directed acyclic graph or combination; According to a prompt word template corresponding to the dependency type, the at least one second operation and maintenance operation and the tool dependency relationship are input into the first model to obtain the operation and maintenance intention corresponding to the at least one operation and maintenance record.

8. The method according to any one of claims 1-7, characterized in that, The method further comprises: In the case that the first information and the second information do not match, the first information indicated by the alternative operation and maintenance process is replaced by the second information to obtain an updated operation and maintenance knowledge base; the first information is: an operation and maintenance operation executed for implementing an operation and maintenance intention corresponding to the operation and maintenance intention description information and a calling mode of an operation and maintenance tool; and the second information is: an operation and maintenance operation executed for implementing the operation and maintenance intention description information and a calling mode of an operation and maintenance tool.

9. The method of claim 4, wherein, Before the operation and maintenance tool description information, the operation and maintenance operation description information and the result description information are input into the second model to generate a calling mode of the target operation and maintenance tool, the method further comprises: According to an application range of an operation and maintenance tool, semantic tags corresponding to the plurality of operation and maintenance tools are determined; the target operation and maintenance tool belongs to the plurality of operation and maintenance tools; Based on the plurality of operation and maintenance tools and the semantic tags corresponding to the plurality of operation and maintenance tools, the constraint decoding rule is constructed.

10. The method of claim 4, wherein, The method further includes: updating the calling mode of the target operation and maintenance tool to the calling mode of the target operation and maintenance tool generated based on the constraint decoding rule.

11. The method according to any one of claims 1-10, characterized in that, The first model is configured to generate an operation and maintenance process according to a prompt word, and obtain an operation and maintenance intention according to an operation and maintenance operation and a tool dependency relationship between operation and maintenance tools implementing the operation and maintenance operation.

12. An operation and maintenance process determination apparatus characterized by comprising: The device includes: an acquisition module configured to acquire operation and maintenance intention description information input by a user; a processing module configured to determine, based on an operation and maintenance knowledge base, a candidate operation and maintenance process associated with the operation and maintenance intention description information, wherein the operation and maintenance knowledge base includes at least one operation and maintenance process, and one operation and maintenance process is configured to indicate operation and maintenance operations performed to implement one operation and maintenance intention and a calling mode of an operation and maintenance tool; the candidate operation and maintenance process is configured to indicate operation and maintenance operations performed to implement an operation and maintenance intention corresponding to the operation and maintenance intention description information and a calling mode of an operation and maintenance tool; the processing module is further configured to generate a prompt word according to the candidate operation and maintenance process and meta information of operation and maintenance tools in an operation and maintenance tool set, wherein the operation and maintenance tools in the operation and maintenance tool set are determined according to the operation and maintenance intention description information; the processing module is further configured to input the prompt word into a model to obtain a target operation and maintenance process, wherein the target operation and maintenance process is configured to indicate a target operation and maintenance operation implemented to implement the operation and maintenance intention description information and a calling mode of a target operation and maintenance tool.

13. The apparatus of claim 12, wherein, The operation and maintenance knowledge base includes at least one operation and maintenance triple, one operation and maintenance triple includes an operation and maintenance intention, at least one first operation and maintenance operation implemented to implement the operation and maintenance intention and an operation order, and at least one first operation and maintenance tool implemented to implement the at least one first operation and maintenance operation and a calling mode of the at least one first operation and maintenance tool, and one operation and maintenance triple corresponds to one operation and maintenance process.

14. The apparatus of claim 13, wherein, The target operation and maintenance process corresponds to a target operation and maintenance triple, and at least one operation and maintenance triple in the operation and maintenance knowledge base includes or does not include the target operation and maintenance triple.

15. The apparatus of claim 14, wherein, The model includes a first model and a second model. The processing module is specifically configured to input the prompt word into the first model to generate operation and maintenance tool description information, operation and maintenance operation description information and result description information implemented to implement the operation and maintenance intention description information; the processing module is further specifically configured to input the operation and maintenance tool description information, the operation and maintenance operation description information and the result description information into the second model to generate a calling mode of the target operation and maintenance tool, wherein the second model is configured to generate the calling mode of the target operation and maintenance tool according to a constraint decoding rule, and the constraint decoding rule includes a plurality of operation and maintenance tools and at least one semantic tag corresponding to the plurality of operation and maintenance tools; one semantic tag corresponds to at least one of the plurality of operation and maintenance tools, one operation and maintenance tool corresponds to at least one of the at least one semantic tag, and a semantic tag is configured to indicate an application range of an operation and maintenance tool.

16. The device of any one of claims 12-15, wherein The processing module is further configured to: acquire at least one operation and maintenance record; one operation and maintenance record comprises: a second operation and maintenance operation, a tool identifier of a first operation and maintenance tool for implementing the second operation and maintenance operation, and a result of calling the first operation and maintenance tool to execute the second operation and maintenance operation; The processing module is further configured to: generate an operation and maintenance process corresponding to the at least one operation and maintenance record according to the at least one operation and maintenance record; the operation and maintenance process corresponding to the at least one operation and maintenance record comprises: an operation and maintenance intention corresponding to the at least one operation and maintenance record, at least one second operation and maintenance operation and an operation sequence corresponding to the at least one operation and maintenance record, and at least one second operation and maintenance tool for implementing the at least one second operation and maintenance operation and a calling mode of the at least one second operation and maintenance tool; The processing module is further configured to: construct the operation and maintenance knowledge base according to the operation and maintenance process corresponding to the at least one operation and maintenance record.

17. The apparatus of claim 16, wherein The processing module is configured to: determine a tool dependency relationship between the at least one second operation and maintenance tool according to an operation dependency relationship between the at least one second operation and maintenance operation; The processing module is further configured to: determine the operation and maintenance intention corresponding to the at least one operation and maintenance record according to the at least one second operation and maintenance operation and the tool dependency relationship; The processing module is further configured to: determine the operation and maintenance process corresponding to the at least one operation and maintenance record according to the operation and maintenance intention corresponding to the at least one operation and maintenance record, the at least one second operation and maintenance operation, the operation dependency relationship, the at least one second operation and maintenance tool, and the tool dependency relationship.

18. The apparatus of claim 17, wherein, The model comprises a first model configured to obtain an operation and maintenance intention according to a tool dependency relationship between an operation and maintenance operation and an operation and maintenance tool for implementing the operation and maintenance operation; The processing module is configured to: determine a dependency type of the tool dependency relationship; the dependency type comprises one of the following: linear, parallel, tree, directed acyclic graph, or combination; The processing module is further configured to: input the at least one second operation and maintenance operation and the tool dependency relationship into the first model according to a prompt word template corresponding to the dependency type, to obtain the operation and maintenance intention corresponding to the at least one operation and maintenance record.

19. The apparatus of any one of claims 12-18, wherein The processing module is further configured to: replace the first information indicated by the alternative operation and maintenance process with second information in a case where the first information and the second information do not match, to obtain an updated operation and maintenance knowledge base; the first information comprises: an operation and maintenance operation executed for implementing an operation and maintenance intention corresponding to the operation and maintenance intention description information and a calling mode of an operation and maintenance tool; and the second information comprises: an operation and maintenance operation executed for implementing the operation and maintenance intention description information and a calling mode of an operation and maintenance tool.

20. The apparatus of claim 16, wherein The processing module is further configured to: determine semantic tags corresponding to the plurality of operation and maintenance tools according to application ranges of the operation and maintenance tools; and the target operation and maintenance tool belongs to the plurality of operation and maintenance tools. The processing module is further configured to construct the constraint decoding rule based on the multiple operation and maintenance tools and the semantic labels corresponding to the multiple operation and maintenance tools.

21. The apparatus of claim 16, wherein, The processing module is further configured to update the calling mode of the target operation and maintenance tool to the calling mode of the target operation and maintenance tool generated based on the constraint decoding rule.

22. The apparatus of any one of claims 12-21, wherein, The first model is configured to generate an operation and maintenance process according to a prompt word, and obtain an operation and maintenance intent according to an operation and maintenance operation and a tool dependency relationship between operation and maintenance tools implementing the operation and maintenance operation.

23. A processor, comprising: The processor includes interface circuitry and control circuitry; the interface circuitry is configured to obtain operation and maintenance intent description information, and the control circuitry is configured to perform the method of any one of claims 1-11 in cooperation with the interface circuitry.

24. A cluster of computing devices, characterized in that, The computing device cluster includes at least one computing device, and each computing device includes a processor and a memory; The processor of the at least one computing device is configured to execute instructions stored in the memory of the at least one computing device, so that the computing device cluster performs the method of any one of claims 1-11.

25. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer instructions; when the computer instructions are executed in a computing device, the computing device performs the method of any one of claims 1-11.

26. A computer program product, characterised in that, When the computer program product is executed in a computing device, the computing device performs the method of any one of claims 1-11.

Images