Interface document construction method, key parameter processing model training method and computing equipment

Abstract

The invention discloses an interface document construction method, a key parameter processing model training method and computing equipment. The interface document construction method comprises the steps that a target data document of a data center is obtained, the target data document comprises a plurality of target data objects, and the target data objects are used for describing configuration information of multiple target facilities in the data center; based on a key parameter processing rule, determining an association relationship among the key configuration information of the multiple target facilities from the target interface data document; performing data construction of the configuration information of the multiple target facilities based on the association relationship among the key configuration information of the multiple target facilities; and based on the constructed data, generating an interface document of the target type.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method for constructing an interface document, a method for training a key parameter processing model, and a computing device. Background Art

[0002] During the development of the upper-level application software of the data center management tool, developers cannot access the real data of the data center stored in the user data center management tool; therefore, it is necessary to simulate the interface document of the user data center management tool based on the target data document of the user data center management tool to support developers in the functional development and debugging of the upper-level application software of the data center management tool.

[0003] The interface document of the user data center management tool includes multiple data objects, and the configuration information of the facilities of the data center described by the multiple data objects may have a certain association relationship. In other words, in multiple data objects, there may be configuration information used to describe the same facility information. The value patterns of these configuration information used to describe the same facility information are the same, but the keys corresponding to the values ​​are not the same; therefore, before constructing the interface document, it is necessary to first determine the association relationship between the configuration information of the multiple facilities according to the target data document, so that in multiple target data objects, the same values ​​can be constructed for these related configuration information, thereby saving the time required for simulation data construction.

[0004] However, when the amount of data information in the target data document or the simulation data document is large, it is relatively slow to manually identify the relationship between the key configuration information of multiple target facilities and construct the related data information. Summary of the invention

[0005] The present application discloses a method for constructing an interface document, a method for training a key parameter processing model, and a computing device. By obtaining the association relationship between key configuration information of multiple target facilities according to pre-configured key parameter processing rules, the problem of slow manual identification of the association relationship between key configuration information of multiple target facilities when the amount of data is large can be solved.

[0006] In a first aspect, an embodiment of the present application provides a method for constructing an interface document, the method being executed by a computing device, and characterized in that the method comprises:

[0007] Acquire a target data document of a data center, wherein the target data document includes a plurality of target data objects, and the target data objects are used to describe configuration information of a plurality of target facilities in the data center;

[0008] Based on the key parameter processing rules, determining the association relationship between the key configuration information of the multiple target facilities from the target interface data document;

[0009] Based on the association relationship between the key configuration information of the multiple target facilities, data construction of the configuration information of the multiple target facilities is performed;

[0010] Based on the constructed data, an interface document of the target type is generated.

[0011] In this way, the association relationship between key configuration information of multiple target facilities can be obtained according to pre-configured key parameter processing rules, thereby solving the problem of slow manual identification of the association relationship between key configuration information of multiple target facilities when the data volume is large.

[0012] In a possible implementation, the key parameter processing rule includes a pre-trained key parameter processing model or a pre-configured rule set; the pre-configured rule set defines at least one regular expression corresponding to each of a plurality of category labels;

[0013] Wherein, the determining of the association relationship between the key configuration information of multiple target facilities from the target interface data document based on the key parameter processing rule includes:

[0014] Matching the regular expression with the configuration information of the multiple target facilities, and determining the configuration information that successfully matches as the key configuration information of the multiple target facilities;

[0015] Based on the patterns of the key configuration information of the multiple target facilities, association relationships between the key configuration information of the multiple target facilities are determined.

[0016] In this implementation, by defining a unified pattern through regular expressions, key configuration information that conforms to a specific regular expression can be quickly filtered out from a large amount of data, ensuring that the key configuration information has a similar pattern, saving the time required to determine the association relationship of the key configuration information, and improving the efficiency of interface document construction.

[0017] In a possible implementation, the method further includes:

[0018] Based on the category label corresponding to the regular expression, determining the category labels corresponding to the key configuration information of the multiple target facilities respectively; the multiple category labels have a hierarchical relationship;

[0019] For any target data object, determining the category label of the target data object according to the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship;

[0020] The category label of the target data object indicates the category of the target facility corresponding to the target data object.

[0021] In this embodiment, since the configuration information of higher-level facilities includes not only the data information of the higher-level facilities themselves, but may also include the data information of lower-level facilities, the category label of a single target data object can be determined based on the different category labels corresponding to the configuration information of a single target facility and the hierarchy between the category labels, thereby obtaining the category of the target facility corresponding to the target data object.

[0022] In a possible implementation manner, determining the category label of the target data object according to the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship includes:

[0023] The highest-level category label among the category labels corresponding to all the key configuration information included in the target data object is determined as the category label of the target data object.

[0024] In this implementation, the category label at the highest level can be directly determined as the category label of the target data object, thereby saving the time required for determining the category label of the target data object and improving the construction efficiency of the interface document.

[0025] In a possible implementation, the configuration information includes a key-value pair;

[0026] The values ​​corresponding to the key configuration information of a plurality of target facilities having an associated relationship have the same pattern.

[0027] In this implementation, the values ​​corresponding to the key configuration information of multiple target facilities having an associated relationship have the same pattern and are all used to describe the same or similar content, for example, they can all be used to describe the location information of a room.

[0028] In a possible implementation, the data construction of the configuration information of the multiple target facilities based on the association relationship between the key configuration information of the multiple target facilities includes:

[0029] Based on the association relationship between the key configuration information of the multiple target facilities, data construction of the key configuration information of the multiple target facilities is performed;

[0030] And, data construction of non-critical configuration information of the various target facilities is performed.

[0031] In this implementation, since the patterns of the values ​​of key configuration information with associated relationships are the same, the same pattern can be used to construct these values ​​with the same pattern when simulating data; and, for the key configuration information of multiple target facilities with corresponding relationships, the same values ​​are constructed for the key configuration information with associated relationships, thereby saving the time required to simulate the configuration information of the management unit and improving the construction efficiency of the interface document.

[0032] In a possible implementation, it is characterized in that the key parameter processing model includes a BERT model.

[0033] Compared with other machine learning models, the BERT model has a bidirectional Transformer architecture that can consider the contextual information of each word when processing it, allowing BERT to better capture the true meaning of words. In addition, the BERT model can be fine-tuned for specific tasks according to user needs, so that good matching results can be achieved even with only a small number of user sample data, thereby improving the efficiency of interface document construction.

[0034] In a possible implementation, matching the regular expression with the configuration information of the multiple target facilities includes:

[0035] The regular expression is matched with the configuration information of the multiple target facilities through BIO notation.

[0036] In this implementation, because the BIO annotation method can clearly define entity boundaries during matching, is easily extended to multiple entity types, and supports efficient and accurate model training and performance evaluation, a more accurate matching result can be obtained through the BIO annotation method, thereby improving the construction efficiency of the interface document.

[0037] In a second aspect, an embodiment of the present application provides a device for constructing an interface document, characterized in that the device includes:

[0038] A data information acquisition module is configured to acquire a target data document of a data center, wherein the target data document includes a plurality of target data objects, and the target data objects are used to describe configuration information of a plurality of target facilities in the data center;

[0039] A key parameter processing module configured to determine, based on key parameter processing rules, associations between key configuration information of multiple target facilities from the target interface data document;

[0040] A data information construction module configured to construct data of the configuration information of the multiple target facilities based on the association relationship between the key configuration information of the multiple target facilities;

[0041] The interface document generation module is configured to generate an interface document of a target type based on the constructed data.

[0042] In this way, the association relationship between key configuration information of multiple target facilities can be obtained according to pre-configured key parameter processing rules, thereby solving the problem of slow manual identification of the association relationship between key configuration information of multiple target facilities when the data volume is large.

[0043] In a possible implementation, the key parameter processing rule includes a pre-trained key parameter processing model or a pre-configured rule set; the pre-configured rule set defines at least one regular expression corresponding to each of a plurality of category labels;

[0044] Among them, the key parameter module is specifically configured as follows:

[0045] Matching the regular expression with the configuration information of the multiple target facilities, and determining the configuration information that successfully matches as the key configuration information of the multiple target facilities;

[0046] Based on the patterns of the key configuration information of the multiple target facilities, association relationships between the key configuration information of the multiple target facilities are determined.

[0047] In this implementation, by defining a unified pattern through regular expressions, key configuration information that conforms to a specific regular expression can be quickly filtered out from a large amount of data, ensuring that the key configuration information has a similar pattern, saving the time required to determine the association relationship of the key configuration information, and improving the efficiency of interface document construction.

[0048] In a possible implementation, the key parameter processing module is further configured to determine the category labels corresponding to the key configuration information of the multiple target facilities based on the category labels corresponding to the regular expressions; the multiple category labels have a hierarchical relationship;

[0049] The device also includes a facility category determination module, which is configured to determine the category label of the target data object for any target data object based on the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship; wherein the category label of the target data object indicates the category of the target facility corresponding to the target data object.

[0050] In this embodiment, since the configuration information of higher-level facilities includes not only the data information of the higher-level facilities themselves, but may also include the data information of lower-level facilities, the category label of a single target data object can be determined based on the different category labels corresponding to the configuration information of a single target facility and the hierarchy between the category labels, thereby obtaining the category of the target facility corresponding to the target data object.

[0051] In a possible implementation, the facility category determination module is specifically configured to determine the highest-level category label among the category labels corresponding to all the key configuration information included in the target data object as the category label of the target data object.

[0052] In this implementation, the category label at the highest level can be directly determined as the category label of the target data object, thereby saving the time required for determining the category label of the target data object and improving the construction efficiency of the interface document.

[0053] In a possible implementation, the configuration information includes a key-value pair;

[0054] The values ​​corresponding to the key configuration information of a plurality of target facilities having an associated relationship have the same pattern.

[0055] In this implementation, the values ​​corresponding to the key configuration information of multiple target facilities having an associated relationship have the same pattern and are all used to describe the same or similar content, for example, they can all be used to describe the location information of a room.

[0056] In a possible implementation manner, the data information construction module is specifically configured as follows:

[0057] Based on the association relationship between the key configuration information of the multiple target facilities, data construction of the key configuration information of the multiple target facilities is performed;

[0058] And, data construction of non-critical configuration information of the various target facilities is performed.

[0059] In this implementation, since the patterns of the values ​​of key configuration information with associated relationships are the same, the same pattern can be used to construct these values ​​with the same pattern when simulating data; and, for the key configuration information of multiple target facilities with corresponding relationships, the same values ​​are constructed for the key configuration information with associated relationships, thereby saving the time required to simulate the configuration information of the management unit and improving the construction efficiency of the interface document.

[0060] In a possible implementation, it is characterized in that the key parameter processing model includes a BERT model.

[0061] Compared with other machine learning models, the BERT model has a bidirectional Transformer architecture that can consider the contextual information of each word when processing it, allowing BERT to better capture the true meaning of words. In addition, the BERT model can be fine-tuned for specific tasks according to user needs, so that good matching results can be achieved even with only a small number of user sample data, thereby improving the efficiency of interface document construction.

[0062] In a possible implementation, the key parameter configuration module is specifically configured to match the regular expression with the configuration information of the multiple target facilities through BIO notation.

[0063] In this implementation, because the BIO annotation method can clearly define entity boundaries during matching, is easily extended to multiple entity types, and supports efficient and accurate model training and performance evaluation, a more accurate matching result can be obtained through the BIO annotation method, thereby improving the construction efficiency of the interface document.

[0064] In a third aspect, a training method for a key parameter processing model is provided. The method comprises:

[0065] Acquire multiple training samples, the training samples include: a sample data document of a data center, a pre-configured sample rule set, and associations between key configuration information of multiple target facilities; wherein the sample data document includes multiple sample data objects, and the sample data objects are used to describe configuration information of multiple target facilities in the data center;

[0066] The model is trained based on the multiple training samples to obtain a key parameter processing model.

[0067] In this method, when receiving a user's data document, the trained key parameter processing model can be used to directly obtain the association relationship between the key configuration information of multiple target facilities from the target data document.

[0068] In a fourth aspect, an embodiment of the present application provides a computing device, comprising a memory and a processor, wherein the memory stores a computer program that can be executed on the processor, and wherein the processor implements the method described in the first aspect or any possible implementation manner of the first aspect when executing the computer program.

[0069] In a fifth aspect, an embodiment of the present application provides a computer storage medium, in which instructions are stored. When the instructions are executed on a computer, the computer executes the method described in the first aspect or any possible implementation of the first aspect.

[0070] In a sixth aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the method described in the first aspect or any possible implementation of the first aspect.

[0071] It can be understood that the beneficial effects of the fourth to sixth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0073] Figure 1 A schematic diagram of the structure of a data center provided as an example in an embodiment of the present application;

[0074] Figure 2 A structural diagram of the corresponding relationship between various facilities provided as an example in the embodiment of the present application;

[0075] Figure 3 A schematic diagram of a JSON data structure corresponding to a target data document provided as an example in an embodiment of the present application;

[0076] Figure 4 One of the flowcharts of a method for constructing an interface document provided in an embodiment of the present application;

[0077] Figure 5 A schematic diagram of a structure of a model provided exemplarily in an embodiment of the present application;

[0078] Figure 6 A flowchart of a training method for a key parameter processing model provided in an embodiment of the present application;

[0079] Figure 7 A second flowchart of a method for constructing an interface document provided in an embodiment of the present application;

[0080] Figure 8 A schematic diagram of the structure of an interface document structure provided in an embodiment of the present application;

[0081] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0082] The technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0083] In order to better understand the embodiments of the present application, some key terms are explained below.

[0084] Data Center: Data Center refers to physical facilities that are specially used to store, manage and process large amounts of data or computing tasks. These facilities usually contain a large number of server cabinets, storage devices, network components and other related infrastructure. Data Center is a core component of modern information technology infrastructure, supporting a variety of businesses such as Internet services, information technology (IT) systems, and cloud computing services.

[0085] Data center management tools: Data center management tools are a collection of software and hardware solutions used to monitor, manage and optimize the physical infrastructure and IT resources of the data center, including but not limited to the Configuration Management Database (CMDB) and Data Center Infrastructure Management (DCIM). CMDB is mainly used to store and manage detailed information on all configuration items in IT resources, which can be hardware, software, documents, services and other IT service-related assets; DCIM is used to monitor, manage and support the information data of the physical infrastructure of the data center, including information data on power, cooling, space, security and other aspects.

[0086] Bidirectional Encoder Representations from Transformers (BERT) is a pre-trained language model based on the Transformer architecture that can use contextual information to generate word representations when processing natural language processing tasks. In addition, because the BERT model can understand the meaning of words in the text bidirectionally, the BERT model can also be used to implement other natural language processing tasks such as text classification, named entity recognition, question-answering systems, text summarization, semantic parsing, etc.

[0087] JS key-value data: JavaScript Object Notation (JSON) is a lightweight data format that is easy for humans to read and write, and easy for machines to parse and generate. It is widely used to transmit data in Web applications and other environments. The basic structure of JSON data includes objects and arrays. Among them, objects are composed of single or multiple key-value pairs; the content of the key is a string; the content of the value can be data of various data types, such as strings, numbers, Boolean values, arrays, objects, NULL, etc. Arrays are composed of lists of values, and the content of the values ​​can be of any type. In JSON data, objects are surrounded by curly brackets "{}", and key-value pairs are separated by commas ","; arrays are surrounded by square brackets "[]", and values ​​are separated by commas ",".

[0088] Pattern: A pattern refers to a certain regularity or characteristic of data that can help identify or classify different objects or content in the data. For example, when it comes to strings, a pattern refers to the format, rules, or template that the string should follow.

[0089] Next, the system architecture or scenario of this application is briefly introduced.

[0090] With the widespread application of cloud computing and big data technology, the data center industry is developing rapidly, and the scale of data centers is getting larger and larger. For example, for a user's data center, the data center may include the following different types of facilities, such as rooms, cabinets, equipment interfaces, etc. Different types of facilities may have different levels; among them, a lower-level facility may be placed or installed with one or more higher-level facilities. For example, Figure 1 and Figure 2 As shown, the user's data center includes three levels of facilities: rooms, cabinets, and equipment interfaces; among them, four cabinets are placed in a single room, and two equipment interfaces are installed on a single cabinet.

[0091] Each facility in the data center has its own configuration information. The configuration information can be represented by JSON data format or other possible data formats. This application mainly uses JSON data as an example for illustrative description. The configuration information of each facility can be regarded as a data object. The configuration information of multiple facilities can be stored as multiple data objects in one data document or in multiple data documents. The data format of the data document can adopt the JSON data format. This application mainly uses the configuration information of one facility as a data object for illustrative description. For the configuration information of a single facility, such as Figure 3As shown, the configuration information of a single facility may include a single or multiple key-value pairs. For a lower-level facility, the configuration information of the lower-level facility may only include its own data information, such as Figure 3 As shown in the target data object "room" in the figure, the configuration information of the room only includes the data information of the room itself; for facilities at a higher level, the configuration information of the higher level facility will not only include the data information of the facility itself, but may also include the data information of other facilities at a lower level, such as Figure 3 As shown in the target data object "cabinet" in the figure, the configuration information of the cabinet includes not only the data information of the cabinet, but also the data information of the room where the cabinet is placed.

[0092] In addition, since there may be a correspondence between various facilities, the configuration information of various facilities may contain corresponding data information, and the corresponding data information may be represented in the JSON data format as the same value but different keys. Figure 2 As shown, there are 4 cabinets in Room A, and each cabinet has 2 equipment interfaces. There is a correspondence between Room A, the 4 cabinets in Room A, and the 8 equipment interfaces; the configuration information of the room, cabinet, and equipment may contain data information about the room. In the configuration information of the facility "Room A", the key corresponding to the value "Room A" indicating the data information of the room is "Description", and in the configuration information of the facility "Cabinet Aa", the key corresponding to the value "Room A" indicating the data information of the room to which the cabinet belongs is "Room"; in the configuration information of the facility "Equipment Interface Aa1", the key corresponding to the value "Rack" indicating the data information of the room to which the equipment interface belongs is "Rack". The above three items of data information all have the same value "Room A", but the keys corresponding to the same value "Room A" are different.

[0093] For data centers with large-scale facilities, users usually need to use dedicated data center management tools to perform daily operation and maintenance management on the configuration information of large-scale facilities. In the process of software development for the upper-level application of the data center management tool, the configuration information of the facility is usually required as data support; for example, when developing the function of visualizing the data of the CMDB system, the configuration information of the facility is required to support the development. In the process of software development, developers cannot connect to the real data in the customer's data center management tool, and thus cannot obtain the real configuration information of the facility from it; therefore, it is necessary to simulate the configuration information of the user's facility according to the target data document of the data center obtained from the user in advance, and simulate the interface service of the data center management tool by, for example, simulating the interface document of the user's data center management tool, so as to support the software development of the data center management tool. Among them, the interface document of the data center management tool refers to the document on how to access the configuration information of the facility stored in the data center management tool through the interface. In the interface document, the key of the facility configuration information may exist in the form of request parameters, request body or response format.

[0094] However, the multiple target data objects included in the target data document obtained from the user in advance may only be partial data objects in the user data center. In other words, there is no corresponding relationship between the multiple facilities corresponding to these target data objects, such as Figure 3 As shown, there are three target data objects in the target data document. The cabinet "XX Computer Room A1-7" installed in the target facility "Equipment Interface B6-6" corresponding to the target data object "Equipment Interface" is not the same as the target facility "XX Computer Room B6-6" corresponding to the target data object "Cabinet". Therefore, there is no corresponding relationship between the target facility "Equipment Interface B6-6" and the standard facility "XX Computer Room B6-6".

[0095] However, the configuration information of these facilities may still have data information describing the same category of facilities, which are represented by the same data value pattern in the JSON data format. For example Figure 3As shown, the value "XX Room 2-A-01" of the data information indicating the room in the target data object "cabinet" has the same pattern as the value "XX Room A1-7" of the data information indicating the room in the target data object "equipment interface". Since the multiple keys in the target data objects corresponding to the configuration information of the same category of facilities are the same, the key corresponding to the value of the configuration information with the same pattern can be determined based on the pattern of the values ​​in the configuration information of multiple facilities that do not have a corresponding relationship, that is, the association relationship between the configuration information of multiple facilities can be determined; wherein, the configuration information of multiple facilities with an association relationship has the same pattern of the value of the configuration information, and is used to describe the same content. Furthermore, when constructing simulation data, the same value can be used for the configuration information of the multiple facilities with a corresponding relationship, and the values ​​of these key configurations with an association relationship can be constructed based on the association relationship between the multiple configuration information, thereby saving the time required for the configuration information of the simulation facilities.

[0096] like Figure 4 As shown, in the process of simulating the configuration information of the user's facilities, the user's target data document is first obtained, wherein the target data document includes multiple target data objects. Secondly, according to the multiple key-value pairs included in the multiple target data objects in the target data document, the content of the data information indicated by the multiple key-value pairs is manually analyzed, and the keys corresponding to the values ​​with the same data format are obtained from the multiple key-value pairs corresponding to the multiple target data objects; then, each key in the target data document is entered into a table, and for the keys corresponding to the values ​​with the same data format, the values ​​corresponding to each key are manually generated in the table; then the values ​​corresponding to other keys are manually generated, and then one or more data documents can be constructed according to a target document of the user, thereby realizing the simulation of the configuration information of the user's facilities. In the case where the user's data document contains the configuration information of large-scale facilities, it is labor-intensive to identify the keys corresponding to the values ​​with the same data format.

[0097] In order to achieve more efficient data relationship identification, an interface document construction method is provided in an embodiment of the present application, which can be executed by a computing device such as a server, etc.; the server can be a rack server, a cabinet server or a blade server, etc., or the server can be a general-purpose server, a GPU server or a DPU server, etc., and a model for key parameter processing can be deployed on the server. The method includes: obtaining a target data document corresponding to a data center, the target data document includes multiple target data objects, and the target data objects are used to describe the configuration information of multiple target facilities in the data center; based on key parameter processing rules, determining the association relationship between the key configuration information of multiple target facilities from the target interface data document; based on the association relationship between the key configuration information of multiple target facilities, performing data construction of the configuration information of multiple target facilities; based on the constructed data, generating an interface document of the target type.

[0098] In this way, the association relationship between key configuration information of multiple target facilities can be obtained according to pre-configured key parameter processing rules, thereby solving the problem of slow manual identification of the association relationship between key configuration information of multiple target facilities when the data volume is large.

[0099] The following first describes in detail the model provided in this application. A model is a function or a set of rules learned from data through an algorithm, which can be used to predict the results or behaviors of unknown data. In short, a model is a computer program that can improve its own performance through training data sets without explicit programming to obtain prediction results.

[0100] Figure 5 A schematic diagram of the architecture of a model provided in an embodiment of the present application, such as Figure 5 As shown, the model provided in the embodiment of the present application may include: a Transformer encoder, a normalization layer, an embedding layer, and an output layer. As described above, the model may be a BERT model. Among them,

[0101] The embedding layer is used to process each word input into the model and add all the processed information together to form the final input representation. The embedding layer processes words through one or more of the following processing methods: the word embedding layer represents the meaning of the word itself; the paragraph embedding layer marks which paragraph the sentence belongs to; and the position embedding layer represents the position information of the word in the sentence.

[0102] The Transformer encoder is the core of the model. The Transformer encoder is composed of multiple identical layers stacked together, each of which contains two sublayers: attention layer and feedforward layer. The attention layer is used to capture the dependencies between different positions in the input representation, compare each position in the input representation with all other positions, calculate the attention weight, and generate new data accordingly. The feedforward layer is a feedforward neural network that performs nonlinear transformations on the output of each position.

[0103] Normalization layers are placed after each sublayer of the Transformer encoder to speed up the training process, and residual connections are applied before each sublayer of the Transformer encoder to help alleviate the vanishing gradient problem.

[0104] The output layer is used to output specific output results of various downstream tasks, including but not limited to text classification, question answering, and other similar language processing tasks.

[0105] When the user wants to train the model, the training samples for training are first obtained, wherein the training samples include: sample data documents corresponding to the data center, pre-configured sample rule sets, and associations between key configuration information of various target facilities.

[0106] For example, the sample data document may contain the following six sample data objects: wherein a single sample data object includes part or all of the data information in the configuration information of a corresponding single facility.

[0107] First data A: {KeyA1: Dongguan computer room 1F-01, KeyA2: computer room, location: Dongguan};

[0108] First data B: {KeyB1: Dongguan computer room 2F-A-01, KeyB2: IT cabinet, KeyB3: A-01, UnitNum: 42};

[0109] First Data C: {KeyC1: Room 01, 1st Floor, Xi'an Laboratory, KeyC2: Cabinet A1-01, BeginU: 1, EndU: 3, UnitNum: 30};

[0110] First data D: {KeyD1: Beijiao computer room 1F01, KeyD2: computer room, location: Zhengzhou};

[0111] First data E: {KeyE1: AI supercomputer room 2F-A-01, KeyE2: IT cabinet, KeyE3: A-01, UnitNum: 21};

[0112] First data F: {KeyF1: Room 01, 1st floor, Xi'an Laboratory, KeyF2: Cabinet A1-01, StartU: 4, EndU: 7, UnitNum: 30}.

[0113] The sample data object used to describe the configuration information of the facility may include one or more keys and values ​​corresponding to the one or more keys. For example, for the first data C, the first data C includes part of the data information in the configuration information of the facility "device interface C". There are 5 keys in the first data C: “KeyC1”, “KeyC2”, “BeginU”, “EndU”, and “UnitNum”; among them, the value corresponding to the key “KeyC1” is the data information of the computer room to which the facility belongs, indicating that the computer room to which the facility “Equipment Interface C” belongs is located in room numbered 01 on the first floor of the Xi’an Laboratory; the value corresponding to the key “KeyC2” is the data information of the cabinet to which the facility belongs, indicating that the cabinet to which the facility “Equipment Interface C” belongs is numbered A1-01; the value corresponding to the key “BeginU” is the data information of the facility itself, indicating that the starting position numbered 1 is where the facility “Equipment Interface C” is installed in the cabinet; the value corresponding to the key “EndU” is the data information of the facility itself, indicating that the ending installation position numbered 3 is where the facility “Equipment Interface C” is installed in the cabinet; the value corresponding to the key “UnitNum” is the data information of the cabinet to which the facility “Equipment Interface C” belongs, indicating that the RU number of the cabinet to which the facility “Equipment Interface C” belongs is 30.

[0114] When a user wants to train a machine learning model, he or she also needs to obtain a sample rule set corresponding to a sample data document. The sample rule set defines at least one regular expression corresponding to each of the multiple category labels. Similar to the different levels between multiple facilities in a data center, multiple sample category labels may also have different or the same levels. For example, there is a sample rule set as follows, in which there are three different levels of sample category labels.

[0115] ROOM: {*computer room, *room, *laboratory*};

[0116] RACK: {UnitNum*, *-*, HeightU*, *-*-*, Unum, *-*-*-*, Rack*, *-*};

[0117] DEVICE (device): {BeginU*, EndU*, StartU*, EndU*, StartU*, *Height, StartU*, *Height}.

[0118] Each sample category label may correspond to a single or multiple regular expressions. The various data information in the configuration information have the same pattern; therefore, by judging whether the various data information included in the configuration information contains a predetermined regular expression and whether the various data information have the same pattern, the key set corresponding to the category label can be determined from the data information; and then the category label corresponding to the configuration information is determined according to the key set corresponding to the category label. Among them, the "*" in the regular expression corresponding to the category label is a wildcard, which means that the content corresponding to the "*" position can be any character except the 0 character; the multiple regular expressions corresponding to the category label are separated by commas ",".

[0119] like Figure 6 As shown, Figure 6 A training method for a key parameter processing model is provided in an embodiment of the present application.

[0120] Step S61, obtaining multiple training samples, the training samples include: a sample data document of the data center, a pre-configured sample rule set, and the association relationship between key configuration information of multiple target facilities; wherein the sample data document includes multiple sample data objects, and the sample data objects are used to describe the configuration information of multiple target facilities in the data center.

[0121] Before training a machine learning model, it is first necessary to obtain sample data, sample rule sets, and the association relationship between key configuration information of multiple target facilities for training; the association relationship between key configuration information of multiple target facilities can be obtained in advance manually from sample data documents.

[0122] Step S63, training the model based on multiple training samples to obtain a key parameter processing model.

[0123] The sample data document and the sample rule set are input into the model, and key configuration information is obtained from the model according to the sample rule set.

[0124] Subsequently, based on the key configuration information and the correlation between the key configuration information of various target facilities, the forward propagation calculation model outputs the correlation relationship identification result according to the key configuration information; and according to the predetermined loss function, the correlation relationship identification result, and the correlation between the key configuration information of various target facilities, the corresponding gradient information is determined and the model parameters of the model are updated according to the gradient information.

[0125] For example, the association relationship between key configuration information of multiple target facilities can be represented by a key set; wherein, multiple keys belonging to the same key set have an association relationship. In other words, when the number of keys included in any key set is greater than 1, any two keys in the arbitrary key set belong to different sample data objects, and the values ​​corresponding to any two core keys have the same pattern.

[0126] The model can identify multiple key-value pairs included in the sample data document according to the regular expression corresponding to the sample category label, and identify the identification key corresponding to the sample category label.

[0127] Then, according to the identified identification key, key set, and loss function, the corresponding gradient information is determined and the model parameters of the model are updated according to the gradient information, thereby completing the training of the model. Subsequently, when receiving the user's target data document, the trained model can be used to directly obtain the key set representing the association relationship between the key configuration information of multiple target facilities from the target data document.

[0128] The technical solutions in the embodiments of the present application are described clearly and completely with specific embodiments below. Obviously, the embodiments described below are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0129] Figure 7 A flowchart of a data relationship processing method provided in an embodiment of the present application.

[0130] See also Figure 7 The method may include but is not limited to part or all of the following steps S71 to S79. The method is executed by a computing device in which a model for key parameter processing is deployed.

[0131] Step S71, obtaining a target data document corresponding to the data center, the target data document including a plurality of target data objects, the target data objects being used to describe configuration information of a plurality of target facilities in the data center.

[0132] The target data document corresponding to the data center can be provided in advance by the user. The multiple target data objects included in the target data document can be multiple configuration information corresponding to multiple facilities in the user data center; wherein the multiple facilities corresponding to the multiple target data objects included in the target data document can be multiple facilities of different categories that have no association relationship.

[0133] Step S73, based on the key parameter processing rules, determine the association relationship between the key configuration information of the multiple target facilities from the target interface data document.

[0134] Among them, the key parameter processing rules include a pre-trained key parameter processing model or a pre-configured rule set; the pre-configured rule set defines at least one regular expression corresponding to each of multiple category labels.

[0135] Exemplarily, the regular expression can be matched with the configuration information of multiple target facilities based on the key parameter processing rule, and the configuration information that successfully matches can be determined as the key configuration information of multiple target facilities. Subsequently, the association relationship between the key configuration information of multiple target facilities is determined based on the pattern of the key configuration information of multiple target facilities. The values ​​corresponding to the key configuration information of multiple target facilities having an association relationship have the same pattern.

[0136] For example, the regular expression can be matched with the key-value pairs in the target data document, and the keys in the successfully matched key-value pairs can be determined as core keys; then the determined core keys can be divided into multiple key sets. As mentioned above, the configuration information includes key-value pairs, and the key configuration information includes core key-value pairs. The association relationship between key configuration information can be represented by a key set; in other words, multiple keys in the same key set have an association relationship, and the values ​​corresponding to multiple keys in the same key set have the same pattern.

[0137] Continuing with the above example, the target data document obtained from the user may include the above six first data A, B, C, D, E, and F, and each first data is a sample data object. Among them, in the target data document, there are 6 key-value pairs that match the regular expression "* computer room" corresponding to the category label "ROOM": key-value pair "KeyA1: Dongguan computer room 1F-01", key-value pair "KeyB1: Dongguan computer room 2F-A-01", key-value pair "KeyC1: Xi'an laboratory 1st floor 01 computer room", key-value pair "KeyD1: North Suburb computer room 1F01", key-value pair "KeyE1: AI supercomputer computer room 2F-A-01", key-value pair "KeyF1: Xi'an laboratory 1st floor 01 computer room". These 6 keys: key "KeyA1", key "KeyB1", key "KeyC1", key "KeyD1", key "KeyF1", key "KeyE1" are determined as core keys, and the values ​​corresponding to these 6 keys have the same pattern "*computer room", and these 6 keys are divided into the key set corresponding to the category label "ROOM".

[0138] Similarly, in the target data document, there are four key-value pairs that match the regular expression "UnitNum*, *-*" corresponding to the category label "RACK": the key-value pair "KeyB3:A-01" and the key-value pair "UnitNum:42" in the first data B, the key-value pair "KeyC2:A1-01 Rack" and the key-value pair "UnitNum:30" in the first data C; the key-value pair "KeyE3:A-01" and the key-value pair "UnitNum:30" in the first data E. itNum:21", the key-value pair "KeyF2:A1-01 cabinet" and the key-value pair "UnitNum:30" in the first data F; these 4 groups of keys: key "KeyB3" and key "UnitNum", key "KeyC2" and key "UnitNum", key "KeyE3" and key "UnitNum", key "KeyF2" and key "UnitNum" are determined as core keys, and these 4 groups of keys are divided into the key set corresponding to the category label "ROOM".

[0139] Similar to the category tag “RACK”, the matching between the regular expression corresponding to the category tag “DEVICE” and the key-value pairs in the target data document will not be described in detail.

[0140] It should be noted that, as shown in the above example, when determining the association relationship between the key configuration information of multiple target facilities, it is also possible to determine the category labels corresponding to the key configuration information of the multiple target facilities based on the category labels corresponding to the regular expressions; the multiple category labels have a hierarchical relationship.

[0141] In a possible implementation, the regular expression may be matched with the configuration information of various target facilities through BIO notation.

[0142] BIO annotation is a commonly used natural language processing annotation method, which is used to represent the part of speech and morphological information of a word in a sentence, and is often used in tasks such as named entity recognition. The BIO annotation method can be used to determine whether there are fields in the key-value pair that match the regular expression. Furthermore, the BIO annotation method can directly annotate the single or multiple key-value pairs to which the fields that match the regular expression belong as one of the three types of labels: B, I, and O. Among them, B indicates that the key-value pair belongs to the beginning of the label, I indicates that the key-value pair belongs to the inside of the label, and O indicates that the key-value pair is not the beginning or inside of any label.

[0143] For example, for the first data B: {KeyB1: Dongguan computer room 2F-A-01, KeyB2: IT cabinet, KeyB3: A-01, UnitNum: 42}, "KeyB1: Dongguan computer room 2F-A-01" should be marked as "I-ROOM"; "KeyB2: IT cabinet" should be marked as "O"; "KeyB3: A-01, UnitNum: 42" should be marked as "I-RACK". Then, the core key corresponding to the label can be obtained according to the labeling result.

[0144] In addition, through the BIO notation method, the key-value pair to which the core key belongs and the position in the sample data object can also be obtained.

[0145] Then, as shown in the first data B in the above example, the key-value pair "KeyB1: Dongguan Computer Room 2F-A-01" is located at the 1st to 14th positions in the first data B, the key-value pair "Key3: A-01" is located at the 29th to 38th positions in the first data B; the key-value pair "Unum: 42" is located at the 41st to 48th positions in the first data B.

[0146] Optionally, in step S75, for any target data object, the category label of the target data object is determined according to the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship, wherein the category label of the target data object indicates the category of the target facility corresponding to the target data object.

[0147] Exemplarily, the category label with the highest level among the category labels corresponding to all the key configuration information included in the target data object is determined as the category label of the target data object.

[0148] Similar to the above content, since the configuration information of higher-level facilities will not only include the data information of the higher-level facilities themselves, but may also include the data information of lower-level facilities; therefore, the category label of a single target data object can be determined based on the different category labels corresponding to the key set to which the single or multiple core keys included in the single target data object belong, as well as the hierarchy between the category labels.

[0149] Different category tags are used to indicate facilities of different categories. According to the category tag of a single target data object, the configuration information indicating that the single target data object is a facility of the same level as the category tag of the single target data object can be indicated.

[0150] Exemplarily, the highest-level category label among the category labels corresponding to the key sets to which all the core keys included in the target data object belong is determined as the category label of the target data object.

[0151] As shown in the above example, a data center can include facilities at three levels: rooms, cabinets, and equipment interfaces, among which rooms are at the lowest level and equipment interfaces are at the highest level; similarly, category tags can include category tags at three levels: "ROOM", "RACK", and "DEVICE", with "ROOM" being the lowest level and "DEVICE" being the highest level. The levels of category tags can correspond one to one with the levels of facilities.

[0152] If a target data object only includes a core key corresponding to the category tag "ROOM", the category tag of the target data object is determined to be the category tag "ROOM".

[0153] For example, as shown in the first data A, the first data A only includes one core key "KeyA1", and the core key "KeyA1" corresponds to the category label "ROOM", then the category label corresponding to the first data A is the category label "ROOM", indicating that the first data A is used to describe the configuration information of the facility "room".

[0154] If a target data object only includes the core key corresponding to the category tag "ROOM" and also includes the core key corresponding to the category tag "RACK", the category tag of the target data object is determined to be the category tag "RACK".

[0155] As shown in the first data B, there are three core keys in the first data B. The category label corresponding to the core key "KeyB1" is the category label "ROOM", and the category label corresponding to the core key "KeyB3" and the core key "UnitNum" is the category label "RACK". It is determined that the category label corresponding to the first data B is the category label "RACK", indicating that the first data B is used to describe the configuration information of the facility "cabinet".

[0156] If a target data object only includes the core key corresponding to the category tag "ROOM", also includes the core key corresponding to the category tag "RACK", and also includes the core key corresponding to the category tag "DEVICE", then the category tag of the target data object is determined to be the category tag "DEVICE".

[0157] For example, as shown in the first data C, there are five core keys in the first data C, the category label corresponding to the core key "KeyC1" is the category label "ROOM", the category label corresponding to the core key "KeyC2" and the core key "UnitNum" is the category label "RACK", and the category label corresponding to the core key "BeginU" and the core key "EndU" is the category label "DEVICE", then it is determined that the category label corresponding to the first data C is the category label "DEVICE", indicating that the first data C is used to describe the configuration information of the facility "equipment interface".

[0158] Step S77: constructing data of configuration information of various target facilities based on the association relationship between key configuration information of various target facilities.

[0159] The data of the configuration information of the constructed multiple target facilities can be described by simulation data objects. The category label of the simulation data object is the same as the category label of the sample data object corresponding to the simulation data object. The number of simulation data objects can be determined according to the pre-determined levels of multiple facilities, the number of facilities at different levels, and the corresponding relationship between multiple facilities. Figure 1 , Figure 2 The data document corresponding to the data center shown may include 52 simulated data objects, that is, there are 52 facilities corresponding to the data center. Among them, the 52 facilities may include 4 facility "rooms"; each facility "room" may include 4 facility "cabinets", and each facility "cabinet" includes 2 facility "equipment interfaces". In the above data center, the 4 facility "cabinets" and 8 facility "equipment interfaces" included in the same facility "room" have a corresponding relationship. Among the data objects corresponding to these 13 facilities, the values ​​corresponding to the 13 core keys of the key set corresponding to the category label "ROOM" are the same. Similarly, for example Figure 2 As shown, a single facility "cabinet" and the two facilities "equipment interfaces" included in the facility "cabinet" have a corresponding relationship. Among the data objects corresponding to the three facilities, the values ​​corresponding to the three core keys of the key set corresponding to the category label "RACK" are the same.

[0160] Exemplarily, for multiple core keys included in the configuration information of multiple target facilities, multiple simulation values ​​corresponding to the multiple core keys are constructed according to the preset mode corresponding to the key set to which the multiple core keys belong; wherein, when the facilities corresponding to the multiple simulation data objects to which the multiple core keys belong in any key set have a corresponding relationship, the same simulation value is constructed for the multiple core keys with associated relationship included in the multiple simulation data objects with the corresponding relationship.

[0161] For the multiple non-core keys included in the configuration information of the multiple target facilities, simulation values ​​corresponding to the multiple non-core keys are constructed according to the preset modes corresponding to the multiple non-core keys.

[0162] For example, the preset mode corresponding to the key set corresponding to the category label "ROOM" may be "XX room - X-XX". Figure 2The 13 facilities shown as corresponding to the facility "Room A" all have core keys belonging to the key set corresponding to the category label "ROOM" in the 13 simulation data objects corresponding to the 13 facilities. For these 13 core keys, according to the preset mode corresponding to the category label "ROOM", the "X" in the preset data mode is filled with the same data, for example, all are constructed as "Beijing Computer Room-A-21", thereby obtaining 13 identical simulation values.

[0163] For a single or multiple non-core keys included in the configuration information of various target facilities, each non-core key may have a corresponding preset mode. Therefore, it is necessary to construct a simulation value corresponding to each non-core key according to the preset mode corresponding to each non-core key.

[0164] Thus, data structure of configuration information of various facilities in the user data center can be realized.

[0165] Step S79, generating an interface document of the target type based on the constructed data.

[0166] The target type interface document refers to a document that describes how to access the configuration information of the facilities stored in the data center management tool through the interface. The key of the configuration information of the target facility may exist in the interface document in the form of request parameters, request body or response format. The type of interface document can be a Representational State Transfer Application Programming Interface (RESTful API) or other types of interface documents, such as Redfish Application Programming Interface (RedfishrApplication Programming Interface, Redf ish API), Simple Object Access Protocol Application Programming Interface (SOAP API), etc.

[0167] After the interface document of the data center management tool is generated, the interface document can be used to support the development of upper-layer software of the user's data center management tool.

[0168] The above mainly introduces the solution provided by the embodiment of the present application from the perspective of the method. In order to realize the above functions, the data processing device includes a hardware structure and / or software module corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the units and algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present application.

[0169] The embodiment of the present application can exemplarily divide the functional modules of the log analysis device according to the above method. For example, the data processing device can include various functional modules corresponding to the various functional divisions, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical functional division. There may be other division methods in actual implementation.

[0170] The embodiment of the present application also provides a data processing device 80. The data processing device includes: a data information acquisition module 81, a key parameter processing module data information construction module 83, and an interface document generation module 84. Among them,

[0171] The data information acquisition module 81 is configured to acquire a target data document corresponding to the data center. The target data document includes a plurality of target data objects. The target data objects are used to describe configuration information of a plurality of target facilities in the data center.

[0172] The key parameter processing module 82 is configured to determine the association relationship between the key configuration information of multiple target facilities from the target interface data document based on the key parameter processing rules.

[0173] The data information construction module 83 is configured to construct data of the configuration information of the various target facilities based on the association relationship between the key configuration information of the various target facilities.

[0174] The interface document generation module 84 is configured to generate an interface document of a target type based on the constructed data.

[0175] Fig. 9 This is a hardware structure diagram of an electronic device provided in an embodiment of the present application. Fig. 9, the electronic device 90 may include a processor 91 and a memory 92. The processor 91 and the memory 92 may communicate with each other; illustratively, the processor 91 and the memory 92 communicate with each other via a communication bus 93.

[0176] The memory 92 is used to store computer executable instructions.

[0177] The processor 91 is used to execute the computer execution instructions stored in the memory 92, so that the processor 91 executes the technical solution shown in the above method embodiment.

[0178] Optionally, the electronic device 90 may further include a communication interface, which may include a transmitter and / or a receiver.

[0179] Optionally, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly implemented as being executed by a hardware processor, or may be implemented by a combination of hardware and software modules in the processor.

[0180] The electronic device provided in the embodiments of the present application can be used to execute the technical solutions shown in any of the above method embodiments. Its implementation principles and technical effects are similar and will not be elaborated here.

[0181] The embodiment of the present application also provides a chip, which is used to execute the version upgrade method executed by any of the above method embodiments. Its implementation principle and technical effect are similar and will not be repeated here.

[0182] The embodiment of the present application also provides a computer-readable storage medium, in which computer instructions are stored. When the computer instructions are executed by a computer, a version upgrade method as performed in any of the above method embodiments is implemented, and its implementation principle and technical effect are similar, which will not be repeated here.

[0183] The embodiment of the present application also provides a computer program product, including a computer program. When the computer program is executed by a computer, it implements the version upgrade method performed by any of the above method embodiments. Its implementation principle and technical effects are similar and will not be repeated here.

[0184] All or part of the steps of the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a readable memory. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned memory (storage medium) includes: read-only memory (ROM), RAM, flash memory, hard disk, solid state drive, magnetic tape, floppy disk, optical disc and any combination thereof.

[0185] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable terminal device to generate a machine, so that the instructions executed by the processing unit of the computer or other programmable terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0186] These computer program instructions can also be loaded onto a computer or other programmable terminal device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0187] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the embodiments of the present application. Thus, if these modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalents, the embodiments of the present application are also intended to include these modifications and variations.

[0188] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0189] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0190] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, rather than limitation, and the above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

[0191] The block diagrams of the devices, apparatuses, equipment, and systems involved in this disclosure are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including," "comprising," "having," and the like are open words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or," and can be used interchangeably therewith, unless the context clearly indicates otherwise.

[0192] It should also be noted that in the apparatus, device and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0193] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for constructing an interface document, the method being executed by a computing device, characterized in that: The method comprises: Acquire a target data document of a data center, wherein the target data document includes a plurality of target data objects, and the target data objects are used to describe configuration information of a plurality of target facilities in the data center; Based on the key parameter processing rules, determining the association relationship between the key configuration information of the multiple target facilities from the target interface data document; Based on the association relationship between the key configuration information of the multiple target facilities, data construction of the configuration information of the multiple target facilities is performed; Based on the constructed data, an interface document of the target type is generated.

2. The method according to claim 1, characterized in that The key parameter processing rules include a pre-trained key parameter processing model or a pre-configured rule set; the pre-configured rule set defines at least one regular expression corresponding to each of multiple category labels; Wherein, the determining of the association relationship between the key configuration information of multiple target facilities from the target interface data document based on the key parameter processing rule includes: Matching the regular expression with the configuration information of the multiple target facilities, and determining the configuration information that successfully matches as the key configuration information of the multiple target facilities; Based on the patterns of the key configuration information of the multiple target facilities, association relationships between the key configuration information of the multiple target facilities are determined.

3. The method according to claim 2, characterized in that The method further comprises: Based on the category label corresponding to the regular expression, determining the category labels corresponding to the key configuration information of the multiple target facilities respectively; the multiple category labels have a hierarchical relationship; For any target data object, determining the category label of the target data object according to the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship; The category label of the target data object indicates the category of the target facility corresponding to the target data object.

4. The method according to claim 3, characterized in that Determining the category label of the target data object according to the category label corresponding to the key configuration information of the target facility included in the target data object and the hierarchical relationship includes: The highest-level category label among the category labels corresponding to all the key configuration information included in the target data object is determined as the category label of the target data object.

5. The method according to claim 2, characterized in that: The configuration information includes key-value pairs; The values ​​corresponding to the key configuration information of a plurality of target facilities having an associated relationship have the same pattern.

6. The method according to claim 5, characterized in that The data construction of the configuration information of the multiple target facilities based on the association relationship between the key configuration information of the multiple target facilities includes: Based on the association relationship between the key configuration information of the multiple target facilities, data construction of the key configuration information of the multiple target facilities is performed; And, data construction of non-critical configuration information of the various target facilities is performed.

7. The method according to claim 2, characterized in that The key parameter processing model includes a BERT model.

8. The method according to claim 2, characterized in that: The matching of the regular expression with the configuration information of the multiple target facilities includes: The regular expression is matched with the configuration information of the multiple target facilities through BIO notation.

9. A training method for a key parameter processing model, characterized in that: include: Acquire multiple training samples, the training samples include: a sample data document of a data center, a pre-configured sample rule set, and associations between key configuration information of multiple target facilities; wherein the sample data document includes multiple sample data objects, and the sample data objects are used to describe configuration information of multiple target facilities in the data center; The model is trained based on the multiple training samples to obtain a key parameter processing model.

10. A computing device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 9 is implemented.

Images