Application metadata processing methods, apparatus, devices and readable storage media

Abstract

The application discloses an application metadata processing method and device, equipment and a readable storage medium, and the method comprises the steps that an initial application link of a data processing execution process is acquired, and metadata of the data processing execution process is acquired; according to the processing characteristic information, the initial application link is disassembled into a processing unit link and a data unit link; according to the dimension transformation feature of the data unit link, the processing unit link and the data unit link are subjected to link reconstruction processing to form a target application link. The application realizes disassembling the initial application link through metadata, obtaining the corresponding processing unit link and data unit link, and according to the dimension transformation feature of the data unit link, recombining and link reconstruction processing are performed on the abstract units in the processing unit link and the data unit link to form the target application link.

Description

Technical Field

[0001] This application relates to the field of big data processing technology, and in particular to an application metadata processing method, apparatus, device and readable storage medium. Background Technology

[0002] An application is a data processing unit that runs on a software platform, and its processing goal is to create a certain service.

[0003] In related technologies, application decomposition methods are mainly divided into the following two categories: The first category is to classify according to module functions, decompose the application into different modules, and then integrate the functions of each module. This method is more suitable for large-scale module division at the system level, but cannot be applied to fine-grained division based on the characteristics of the data itself. The second category is to classify functions based on code module parsing. This method is based on the analysis system of the impact of application resource changes formed by reverse parsing and decomposition of software code. This method is suitable for processing processes with full source code, but cannot be applied in commercial environments where the source code cannot be compiled and the module scope is difficult to manage, and the processing efficiency cannot be guaranteed.

[0004] Furthermore, in related technologies, the matrix dimensionality features of data units are widely used in matrix transformation processes, including image processing and image recognition, for dimensionality reduction and feature extraction. A common approach is to transform the matrix dimensionality features of data units according to a certain encoding or dimensional space mapping, and then further process and analyze based on the known transformation features to form the final output. However, such methods cannot process summary text objects that describe data, such as application metadata.

[0005] In summary, there are application metadata processing methods in related technologies that lack universality. Summary of the Invention

[0006] In view of this, this application provides an application metadata processing method, apparatus, device, and readable storage medium, aiming to improve the versatility of the application metadata processing method.

[0007] To achieve the above objectives, this application provides an application metadata processing method, which includes the following steps:

[0008] Obtain the initial application chain of the data processing execution process;

[0009] Extract the processing feature information of the initial application link, and decompose the initial application link into a processing unit link and a data unit link based on the processing feature information; wherein, the processing feature information includes metadata for describing the processing process of the initial application link;

[0010] Based on the dimensional transformation characteristics of the data unit link, the processing unit link and the data unit link are reconstructed to form the target application link.

[0011] For example, the step of extracting the processing feature information of the initial application link and decomposing the initial application link into processing unit links and data unit links based on the processing feature information includes:

[0012] Determine the application processing procedure of the initial application link and extract the processing feature information of the application processing procedure;

[0013] Based on the processing feature information, the processing actions involved in the initial application link are determined, and based on the processing actions, the initial application link is decomposed into object units, and the object units are reconstructed into processing unit links.

[0014] Based on the processing feature information, the input matrix and output matrix corresponding to the object unit in the processing unit link are generated, and the input matrix and the output matrix are arranged according to a preset rule to form a data unit link.

[0015] For example, the step of generating the input matrix and output matrix corresponding to the object unit in the processing unit link based on the processing feature information, and arranging the input matrix and the output matrix according to a preset rule to form a data unit link includes:

[0016] Based on the processing feature information, determine the data unit features of the object unit in the processing unit link;

[0017] Based on the characteristics of the data unit, the input matrix and output matrix corresponding to the object unit are generated, and the matrix dimensions of the input matrix and the output matrix are marked.

[0018] Based on the matrix dimension, the input matrix and the output matrix are arranged according to a preset rule to form a data unit link.

[0019] For example, after the steps of generating the input matrix and output matrix corresponding to the object unit based on the data unit characteristics, and marking the matrix dimensions of the input matrix and the output matrix, the method further includes:

[0020] The link components in the processing unit link and the data unit link are characterized;

[0021] The link component unit includes the data processing unit and the data conversion unit corresponding to the input matrix and the output matrix. When the data conversion unit performs feature labeling, the input matrix and the output matrix are labeled as a whole.

[0022] For example, the step of reconstructing the processing unit link and the data unit link based on the dimensional transformation characteristics of the data unit link to form the target application link includes:

[0023] Based on the dimensional transformation characteristics of the data unit link, compare the matrix dimensional transformation information between each adjacent data object in the processing unit link and the data unit link.

[0024] Based on the transformation information, adjacent data objects that have a dimensional transformation relationship are merged to obtain an abstract unit;

[0025] The abstract unit is reconstructed to form the target application link.

[0026] For example, the target application link includes a first target application link and a second target application link, and the step of performing link reconstruction processing on the abstract unit to form the target application link includes:

[0027] The abstract units are marked independently, and the links where the abstract units are located are recombined and reconstructed to form the first target application link;

[0028] The abstract unit is individually reconstructed into a complete link, and the complete link is used as the second target application link.

[0029] For example, the step of merging adjacent data objects with dimensional transformation associations according to the transformation information to obtain an abstract unit includes:

[0030] Based on the transformation information, determine the dimension transformation expression of the adjacent data objects;

[0031] If there is only one value of zero in the dimension transformation expression, it is determined that the adjacent data objects have the transformation information, and they are merged to obtain an abstract unit.

[0032] For example, to achieve the above objectives, this application also provides an application metadata processing apparatus, the apparatus comprising:

[0033] The acquisition module is used to acquire the initial application chain of the data processing execution process;

[0034] The decomposition module is used to extract the processing feature information of the initial application link and decompose the initial application link into processing unit links and data unit links based on the processing feature information; wherein, the processing feature information includes metadata for describing the processing process of the initial application link;

[0035] The processing module is used to perform link reconstruction processing on the processing unit link and the data unit link according to the dimensional transformation characteristics of the data unit link, so as to form the target application link.

[0036] For example, to achieve the above objectives, this application also provides an application metadata processing device, the device comprising: a memory, a processor, and an application metadata processing program stored on the memory and executable on the processor, the application metadata processing program being configured to implement the steps of the application metadata processing method as described above.

[0037] For example, to achieve the above objectives, this application also provides a computer-readable storage medium storing an application metadata processing program, which, when executed by a processor, implements the steps of the application metadata processing method described above.

[0038] In related technologies, methods for application decomposition and matrix transformation of data unit matrix dimension features are not applicable to summary text objects describing data, such as application metadata. This results in a lack of universality in application metadata methods. In contrast, this application obtains the initial application link of the data processing execution process; extracts the processing feature information of the initial application link; and decomposes the initial application link into processing unit links and data unit links based on the processing feature information. The processing feature information includes metadata describing the processing of the initial application link. Based on the dimension transformation features of the data unit links, the processing unit links and the data unit links are reconstructed to form the target application link. In other words, by decomposing the initial application link based on its processing characteristics, corresponding processing unit links and data unit links are obtained. Then, based on the dimensional transformation characteristics of the data unit links, the abstract units in the processing unit links and data unit links can be reconstructed. This means that based on metadata, the initial application link and the decomposed processing unit links and data unit links can be processed accordingly. This enables the combination and verification of the characteristics of the corresponding links based on application metadata. Compared with related technologies, this method is more suitable for application metadata scenarios, thus improving the versatility of the data processing method. Attached Figure Description

[0039] Figure 1This is a flowchart illustrating the first embodiment of the metadata processing method applied in this application;

[0040] Figure 2 A detailed flowchart illustrating step S120 of the first embodiment of the metadata processing method applied in this application;

[0041] Figure 3 A detailed flowchart of step S130 of the first embodiment of the metadata processing method applied in this application;

[0042] Figure 4 This is a schematic diagram of the hardware operating environment involved in the embodiments of this application.

[0043] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0044] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0045] This application provides an application metadata processing method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the metadata processing method applied in this application.

[0046] This application provides an embodiment of an application metadata processing method. It should be noted that although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order. For ease of description, the execution entity descriptions of the various steps of the application metadata processing method are omitted below. The application metadata processing method includes:

[0047] Step S110: Obtain the initial application link of the data processing execution process;

[0048] In this embodiment, the relevant application decomposition methods are not applicable to the fine-grained decomposition and management of application metadata. These application decomposition methods mainly include two categories: Method 1: Classification by module function. This method enables broad category classification based on module function and is more suitable for large-scale system-level module division, but it is not suitable for fine-grained division based on the characteristics of the data itself. Method 2: Decomposition based on the parsing and implementation units of software source code. This method is suitable for processing processes with full source code, but it is not suitable for commercial environments where source code cannot be compiled, module scope is difficult to manage, and processing efficiency cannot be guaranteed.

[0049] Meanwhile, the relevant data feature reorganization methods are not applicable to feature extraction in the metadata management dimension of applications. These data feature reorganization methods mainly include three types: Method 1: Processing unit reorganization based on matrix encoding transformation. This method is suitable for mathematical transformations using specific numbers as feature matrices, but cannot process summary text objects describing application metadata. Method 2: Processing unit reorganization based on matrix dimension space transformation. Similar to Method 1, the basic principle of matrix dimension space transformation is to map vectors to another dimension according to a certain mapping relationship to achieve data reorganization. Multiple vectors are transformed to other spaces according to fixed rules to form new data features. This method is more suitable for multi-dimensional vector combinations represented by numbers, but difficult to use with one-dimensional metadata feature data. Method 3: Reorganization based on the combination of multiple features to form the latest feature to achieve the processing target. This method reorganizes data features by performing logical transformations under a fixed physical meaning mapping on the combination of multiple features. It is suitable for scenarios where the physical meaning of multiple object entities is clear, but difficult to use with one-dimensional metadata feature data.

[0050] In summary, in order to form a universal application metadata decomposition method, this embodiment proposes a custom universal application metadata decomposition method and a method for unit reorganization based on data unit matrix dimension transformation. Ultimately, it forms a unified input and output of data unit features, providing an effective processing method for the functional modular management of application running entities. It is more suitable for the decomposition and data feature reorganization process of metadata of different types of application entities in commercial environments, and has universality, practicality and efficiency.

[0051] This embodiment proposes an application metadata processing method. The application link involved in this application refers to the data processing execution process involved in the data processing application and the data unit matrix before and after the operation of each functional block. The application link includes an initial application link and a target application link. The initial application link is the link to be processed by the application metadata of this embodiment, and the target application link is the link obtained after processing by the application metadata of this embodiment.

[0052] Therefore, in order to achieve the effect of dissecting the application chain through metadata, in this embodiment, the initial application chain of the data processing execution process is obtained, as well as the metadata of the data processing execution process is obtained.

[0053] In this embodiment, after the application link is dissected and broken down using metadata, it can be divided into a processing unit link and a data unit link.

[0054] Step S120: Extract the processing feature information of the initial application link, and decompose the initial application link into a processing unit link and a data unit link according to the processing feature information; wherein, the processing feature information includes metadata for describing the processing process of the initial application link;

[0055] In this embodiment, the processing unit link refers to the link constructed by the data processing execution process involved in the data processing application.

[0056] In this embodiment, the data unit link refers to the link formed by the data unit matrix before and after each functional block of the data processing application runs.

[0057] For ease of explanation, we will use matrix dimensions to represent data units, such as Jmn, which is denoted as a matrix data unit object with m rows and n columns. Here, m and n are the dimensional features of the data unit. Dimension transformation refers to the change of these two values, both of which are greater than or equal to 1.

[0058] In this embodiment, the basic idea of ​​breaking down the initial application chain is to describe the application processing process based on human cognition. Each processing process is described at the smallest understandable functional unit, which will form code segments / functional modules during the development process.

[0059] It should be noted that the description of the application processing process is mainly based on processing feature information. This processing feature information includes metadata used to describe the processing process of the initial application link. That is, this processing feature information is mainly used to describe the initial application link and locate the data processing process of the initial application link. Based on the metadata, each processing process can be clearly located and decomposed, thereby achieving the effect of decomposing the initial application link with the smallest functional unit.

[0060] In this embodiment, taking the implementation process of AIOPS log anomaly prediction application as an example, the log line selection bag-of-words statistics is used as the basic method for text to be converted into a matrix. The processing unit chain can be determined as: CH{preprocessing, time window selection, keyword statistics, data annotation, feature matrix generation, matrix input pre-conversion, AI algorithm calculation, calculation result business conversion, triggering OPS operation, operation execution};

[0061] In this process, CH{preprocessing, time window selection, keyword statistics, data labeling, feature matrix generation, matrix input pre-transformation, AI algorithm calculation, calculation result business conversion, triggering OPS operation, operation execution} is labeled as CH{E1, E2, E3……E10}. The data units for input and output of each processing step are labeled as DH{J1, J2, J3……J9}. The execution process can be represented as CZ{J1, E1, J2, E2, J3, E3……J9, E9, E10}. Here, CH is the processing unit link, DH is the data unit link, and CZ is the initial application link.

[0062] Step S130: Based on the dimensional transformation characteristics of the data unit link, perform link reconstruction processing on the processing unit link and the data unit link to form the target application link.

[0063] In this embodiment, based on the dimensional transformation characteristics of the data unit link, the processing unit link and the data unit link are reconstructed to form the target application link.

[0064] The dimension transformation feature is a pre-defined rule for the dimension transformation of data units. This rule mainly uses matrix dimensions to represent data units. For example, Jmn is a matrix data unit object with m rows and n columns. Here, m and n are the dimension features of the data unit. Dimension transformation refers to the change of these two values, both of which are greater than or equal to 1.

[0065] The abstract unit is the unit containing matrix dimension information in the processing unit link and data unit link obtained after decomposing the initial application link with metadata, which is used to represent each processing function and processing result.

[0066] In this embodiment, when decomposing the initial application link based on metadata, the decomposition is performed at the smallest functional unit level. At the same time, combined with the corresponding matrix dimensions, multiple processing processes with corresponding relationships can be identified in the initial application link. For example, the input and output of multiple processing processes are simultaneously affected by the matrix dimension transformations involved in multiple processing processes. Therefore, in order to ensure the decomposition effect, after obtaining the processing unit link and data unit link, the corresponding abstract units need to be recombined and the link reconstructed. That is, the corresponding abstract units with matrix dimension transformation relationships are processed, so that the corresponding target application link is obtained after the processing is completed.

[0067] In related technologies, methods for application decomposition and matrix transformation of data unit matrix dimension features are not applicable to summary text objects describing data, such as application metadata. This results in a lack of universality in application metadata methods. In contrast, this application obtains the initial application link of the data processing execution process; extracts the processing feature information of the initial application link; and decomposes the initial application link into processing unit links and data unit links based on the processing feature information. The processing feature information includes metadata describing the processing of the initial application link. Based on the dimension transformation features of the data unit links, the processing unit links and the data unit links are reconstructed to form the target application link. In other words, by decomposing the initial application link based on its processing characteristics, corresponding processing unit links and data unit links are obtained. Then, based on the dimensional transformation characteristics of the data unit links, the abstract units in the processing unit links and data unit links can be reconstructed. This means that based on metadata, the initial application link and the decomposed processing unit links and data unit links can be processed accordingly. This enables the combination and verification of the characteristics of the corresponding links based on application metadata. Compared with related technologies, this method is more suitable for application metadata scenarios, thus improving the versatility of the data processing method.

[0068] For example, based on the first embodiment of the metadata processing method of this application described above, another embodiment is proposed. In this embodiment, reference is made to... Figure 2 The step of decomposing the initial application link into processing unit links and data unit links based on the processing feature information includes:

[0069] Step S210: Determine the application processing procedure of the initial application link and extract the processing feature information of the application processing procedure;

[0070] In this embodiment, the initial application link corresponds to the corresponding data processing process. This processing process has certain processing sequence, processing content, and processing methods used in the processing process, as well as other processing feature information related to data processing. This processing feature information can be represented in the form of metadata. The process of extracting this processing feature information is the process of analyzing and decomposing the initial application link, and the processing process and processing features of the initial application link are located and described in detail.

[0071] Step S220: Based on the processing feature information, determine the processing actions involved in the initial application link, and based on the processing actions, decompose the initial application link into object units, and reconstruct the object units into processing unit links.

[0072] In this embodiment, when the initial application link is decomposed and the corresponding processing unit link is constructed based on the processing feature information, that is, based on the corresponding metadata, the initial application link is divided into fine-grained parts according to the data's own characteristics, so as to ensure that each unit in the processing unit link obtained during the decomposition is the smallest functional unit.

[0073] It should be noted that the main purpose of dividing the initial application chain in this way is to divide it into fine-grained units based on the characteristics of the data itself. This allows for arbitrary combination of the smallest units, thereby enabling more flexibility and precision in the subsequent integration of functional modules. Compared with related technologies, the functional modules obtained are of better quality.

[0074] To achieve the above objectives, the processing actions involved in the initial application chain are determined through this metadata. Based on these processing actions, the initial application chain is decomposed, and the corresponding data from the decomposition is used to construct a processing unit chain. Here, the processing action refers to the different processing procedures involved in the initial application chain after detailed data division, such as data preprocessing, data filtering, and data transformation, which can be determined based on the metadata. Thus, the initial application chain can be decomposed into object units based on these processing actions, and these object units can be reorganized and arranged to construct a processing unit chain. Each unit in the processing unit chain corresponds to a processing action (process).

[0075] Taking the AIOPS log anomaly prediction application implementation process mentioned in the above embodiments as an example, the processing unit link can be determined as: CH{preprocessing, time window selection, keyword statistics, data labeling, feature matrix generation, matrix input pre-transformation, AI algorithm calculation, calculation result business conversion, triggering OPS operation, operation execution}. In this processing unit link, the above units all correspond to data processing actions. These processing actions have a certain order. For example, it is necessary to first perform corresponding preprocessing on the data (noise reduction or removal of some error data), and then select some preprocessed data through a time window to narrow the corresponding data processing range. Then, perform corresponding statistics and data labeling on the data within the window, so as to realize the subsequent corresponding data processing process. The specific data processing process varies in different functional implementation scenarios. It is necessary to determine the processing actions included in the initial application link in combination with the actual application scenario. No specific limitation is made here.

[0076] Step S230: Based on the processing feature information, generate the input matrix and output matrix corresponding to the object unit in the processing unit link, and arrange the input matrix and the output matrix according to a preset rule to form a data unit link.

[0077] In this embodiment, after the corresponding processing unit links are obtained by disassembly, the processing unit links need to be described based on the characteristics of the data itself and combined with metadata. In this way, by describing the data characteristics, the description content is used as the unit of the link to form a data unit link.

[0078] In describing data characteristics, it is necessary to record the characteristic information of each processing action based on the metadata and the characteristic information of each processing action in the processing unit link, and describe the characteristic information accordingly. At this time, the method of combining metadata can be used to accurately describe each processing process. This description mainly describes the input matrix and output matrix corresponding to each processing action, and arranges the input matrix and output matrix according to the preset rules to form the data unit link.

[0079] The preset rule arrangement mainly arranges the input and output matrices according to the order of certain processing steps. In addition, it is also necessary to consider the dimensional transformation relationship between the input and output matrices and sort them accordingly, and arrange the data with certain dimensional transformation relationships in a combined manner.

[0080] For example, the step of recording and describing the input matrix and output matrix corresponding to the processing action based on the processing feature information and the feature information of each processing action in the processing unit link, and arranging the input matrix and the output matrix according to a preset rule to form a data unit link, includes:

[0081] Step a: Determine the data unit characteristics of the object unit in the processing unit link based on the processing feature information;

[0082] Step b: Based on the characteristics of the data unit, generate the input matrix and output matrix corresponding to the object unit, and mark the matrix dimensions of the input matrix and the output matrix;

[0083] Step c: Arrange the input matrix and the output matrix according to the matrix dimension to form a data unit link.

[0084] In this embodiment, a corresponding data unit link is generated based on the metadata and feature information of each processing action in the processing unit link. In this process, the processing procedure in the processing unit link is determined by parsing the feature information of each processing action, and the data unit feature formed by the processing procedure is determined at the same time. The data unit feature is the feature of each component unit of the processing unit link CH. Then, the input matrix and output matrix corresponding to the processing action can be recorded and described according to the data unit feature. The above two matrices are the input and output corresponding to a certain processing procedure and are represented in matrix form. At this time, in order to comprehensively consider the relationship between the input matrix and the output matrix, after determining the input matrix and the output matrix, the matrix dimension of the input matrix and the output matrix is ​​marked. Then, the input matrix and the output matrix can be arranged according to the matrix dimension according to the preset rules to form a data unit link.

[0085] It should be noted that in this embodiment, the matrix dimension feature is introduced. The transformation features of this matrix dimension can be used to construct corresponding data unit links. Compared to related technologies that use matrix encoding-based transformations to reorganize processing units, or matrix dimension space transformations to reorganize processing units, the basic principle of matrix dimension space transformation is to map vectors to another dimension according to a certain mapping relationship to achieve data reorganization. Other methods, such as combining multiple features to form the latest feature to achieve the processing target, or performing logical transformations under fixed physical meaning mappings on combinations of multiple features, all involve transforming multiple vectors to other spaces according to fixed rules to form new data features. These methods are more suitable for multi-dimensional vector combinations represented by numbers and are difficult to use for metadata features described in one dimension. However, this embodiment considers the dimensional transformation features between the input and output matrices to achieve data feature reorganization, making it suitable for the decomposition and data feature reorganization processes of metadata for different types of application entities.

[0086] In this embodiment, the specific process for arranging and forming data unit links is as follows: Based on the data unit characteristics formed by the processing of the processing unit links, the matrix dimensions of the data objects are marked. Combining the characteristics of the CH algorithm, let the dimension of J1 be (a, z), and let J1az represent that J1 is a matrix of a rows and z columns. Then, after E1 preprocessing, J2az; after E2 time window selection, J3bz (b is a value less than a); after E3 keyword statistics, J4bc (c is the number of keywords); after E4 data annotation, J5b(c+1); after E5 feature matrix generation, J6de; after E6 matrix input pre-transformation, J7fg; after E7 AI algorithm calculation, J8hi; and after E8 business conversion of the calculation result, J9jk. The final data unit link DH is {J1az, J2az, J3bz, J4bc, J5b(c+1), J6de, J7fg, J8hi, J9jk}. The application link CZ is CZ{J1az, E1, J2az, E2, J3bz, E3, J4bc, E4, J5b(c+1), E5, J6de, E6, J7fg, E7, J8hi, E8, J9jk, E9, E10}.

[0087] For example, after the step of arranging the input matrix and the output matrix according to the matrix dimension according to a preset rule to form a data unit link, the method includes:

[0088] Step d: Characterize the link components in the processing unit link and the data unit link;

[0089] The link component unit includes the data processing unit and the data conversion unit corresponding to the input matrix and the output matrix. When the data conversion unit performs feature labeling, the input matrix and the output matrix are labeled as a whole.

[0090] In the above embodiments, CH{preprocessing, time window selection, keyword statistics, data labeling, feature matrix generation, matrix input pre-transformation, AI algorithm calculation, calculation result business conversion, triggering OPS operation, operation execution} is labeled as CH{E1, E2, E3……E10}, and the data units input and output of each processing process are labeled as DH{J1, J2, J3……J9}. The execution process can be represented as CZ{J1, E1, J2, E2, J3, E3……J9, E9, E10}. The above process is the process of labeling the link components in the corresponding processing unit link and data unit link.

[0091] The link component unit includes a data processing unit corresponding to the processing action and a data conversion unit corresponding to the input matrix and the output matrix. When the data conversion unit is marked, the input matrix and the output matrix are marked as a whole. This ensures that the corresponding target application link can be more accurately divided in the future.

[0092] In this embodiment, based on the processing feature information, the processing actions involved in the initial application link are determined, and based on the processing actions, the initial application link is decomposed and a processing unit link is constructed. Based on the processing feature information and the feature information of each processing action in the processing unit link, the input matrix and output matrix corresponding to the processing action are recorded and described, and the input matrix and output matrix are arranged according to a preset rule to form a data unit link. In other words, by parsing metadata, the initial application link is analyzed to obtain the corresponding processing actions, and based on these processing actions, a processing unit link can be obtained. Furthermore, based on this processing unit link, the corresponding input matrix and output matrix can be recorded and described, and further, after arranging them, the corresponding data unit link can be obtained. This achieves a more detailed decomposition of the initial application link into processing unit links from the perspective of metadata.

[0093] For example, based on the first embodiment of the metadata processing method of this application described above, another embodiment is proposed. In this embodiment, reference is made to... Figure 3 The step of reconstructing the processing unit link and the data unit link based on the dimensional transformation characteristics of the data unit link to form the target application link includes:

[0094] Step S310: Based on the dimensional transformation characteristics of the data unit link, compare the matrix dimensional transformation information between each adjacent data object in the processing unit link and the data unit link.

[0095] In this embodiment, based on the dimensional transformation characteristics of the data unit link, the matrix dimensional transformation association (dimensional transformation characteristics) between each adjacent data object in the processing unit link and the data unit link can be determined. The dimensional transformation characteristics are the transformation information of the baseline rules and specific matrix dimensional transformation associations used to determine the matrix dimensional transformation association between each adjacent data object. As long as it is determined that the matrix dimensional transformation between each adjacent data object conforms to the dimensional transformation characteristics, the matrix dimensional transformation between the adjacent data objects can be associated.

[0096] Therefore, based on this dimensional transformation characteristic, by comparing each adjacent data object in the processing unit link and the data unit link, it can be determined whether there is a matrix dimensional transformation relationship between adjacent data objects.

[0097] Step S320: Based on the transformation information, merge adjacent data objects that have a dimensional transformation relationship to obtain an abstract unit;

[0098] In this embodiment, by comparing the matrix dimensions of adjacent data objects and merging the processing steps with dimension transformation relationships according to the corresponding transformation information, a final abstract unit is formed. The specific steps to obtain the abstract unit are as follows: The second-layer data unit layer link C formed by the processing step S2-2 is updated to {J1az, J2az, J3bz, J4bc, J5b(c+1), J6de, J7fg, J8hi, J9jk}. According to (Formula 1), the three datasets J2az, J3bz, J4bc, and J5b(c+1) are related. According to the application chain of S2-1, CZ{J1az, E1, J2az, E2, J3bz, E3, J4bc, E4, J5b(c+1), E5, J6de, E6, J7fg, E7, J8hi, E8, J9jk, E9, E10}, the execution process with J2az as input and J5b(c+1) as output includes E2, E3, and E4. Therefore, the three processes are merged to form a new application chain: CZ1{J1az, E1, J2az, Pk, J5b(c+1), E5, J6de, E6, J7fg, E7, J8hi, E8, J9jk, E9, E10}, where Pk includes the processing unit chain {E2, E3, E4} and the data unit chain {J2az, J2bz, J4bc, J5b(c+1)}.

[0099] For example, the step of merging adjacent data objects containing the transformation information to obtain an abstract unit includes:

[0100] Step e: Determine the dimensional transformation expression of the adjacent data objects based on the transformation information;

[0101] Step f: If there is only one value of zero in the dimension transformation expression, it is determined that the adjacent data objects have the transformation information, and they are merged to obtain an abstract unit.

[0102] In this embodiment, when merging adjacent data objects that have a matrix dimension transformation relationship, it is necessary to first determine whether there is a relationship between the adjacent data objects. At this time, based on the dimension transformation characteristics, the dimension transformation expression of the adjacent data objects is determined, and based on the situation of the dimension transformation expression, it can be determined whether there is a matrix dimension transformation relationship between the adjacent data objects.

[0103] In this context, the adjacent data objects refer to data that are processed in a sequential order during the data processing process. Based on this processing order, the adjacent data can be considered as adjacent data objects.

[0104] Specifically, based on the established link, if there is only one value of zero in the dimension transformation expression, it is determined that the adjacent data objects have the transformation information. Taking two matrices Eab and Ecd as an example, the dimension transformation expressions ab and cd are judged. If there is only one value of 0 in the above two dimension transformation expressions, it can be determined that the adjacent data objects have a matrix dimension transformation association.

[0105] Step S330: Perform link reconstruction processing on the abstract unit to form the target application link.

[0106] In this embodiment, the abstract units are recombined and the links are reconstructed to form the corresponding target application links. In this process, the adjacent data objects with dimensional transformation relationships are mainly considered, and their corresponding abstract units are recombined and the links are reconstructed to generate multiple accurate application links. That is, after the above decomposition and reorganization, the target application links will have at least two application links.

[0107] For example, the target application link includes a first target application link and a second target application link. The step of recombining and reconstructing the abstract unit to form the target application link includes:

[0108] Step g: The abstract unit is marked independently, and the links where the abstract unit is located are recombined and reconstructed to form the first target application link;

[0109] Step h: Reconstruct the abstract unit into a complete link, and use the complete link as the second target application link.

[0110] In this embodiment, the target application link includes a first target application link and a second target application link. The first target application link is obtained by independently marking the abstract unit and then recombining and reconstructing the link where the abstract unit is located. The second target application link is obtained by directly reconstructing the abstract unit into a complete link.

[0111] The specific link reconstruction process is as follows: the merged abstract unit is marked independently, and the original link is regarded as a processing unit process of the processing unit link. The new link is extended to form a complete link.

[0112] For example, in the above embodiments, the input and output data units of Pk are recorded as Pki and Pko, where Pki = J2az and Pko = J5b(c+1). Based on this PK, a corresponding link 1 (first target link) can be generated: Application link: CZ1{J1az, E1, Pki, Pk, Pko, E5, J6de, E6, J7fg, E7, J8hi, E8, J9jk, E9, E10}, its processing unit link can be recorded as CH1{E1, Pk, E5, E6, E7, E8, E9, E10}, and its data unit link can be recorded as DH1{J1az, Pki, Pko, J6de, J7fg, J8hi, J9jk}.

[0113] Furthermore, a second target link can be generated: Application link: Pk{J2az, E2, J3bz, E3, J4bc, E4, J5b(c+1)}, its processing unit link can be denoted as {E2, E3, E4}, and its data unit link can be denoted as {J2az, J3bz, J4bc, J5b(c+1)}.

[0114] In this embodiment, based on the dimensional transformation characteristics of the data unit link, the matrix dimensional transformation relationships between adjacent data objects in the processing unit link and the data unit link are compared; adjacent data objects with the transformation information are merged to obtain an abstract unit; the abstract unit is then recombined and the link is reconstructed to form the target application link. In other words, after decomposing the initial application link to obtain the corresponding data unit link, the matrix dimensional transformation relationships between adjacent data objects in the processing unit link and the data unit link are compared based on the dimensional transformation characteristics of the data unit link. This allows adjacent data objects with corresponding dimensional transformation relationships to be merged to obtain the corresponding abstract unit. Furthermore, the target application link can be obtained based on this abstract unit. Through this process, the initial application link can be analyzed and reorganized in greater detail, while also ensuring the universality of the application metadata processing method, enabling its application to different scenarios.

[0115] Furthermore, this application also provides an application metadata processing apparatus, which includes:

[0116] The acquisition module is used to acquire the initial application chain of the data processing execution process;

[0117] The decomposition module is used to extract the processing feature information of the initial application link and decompose the initial application link into processing unit links and data unit links based on the processing feature information; wherein, the processing feature information includes metadata for describing the processing process of the initial application link;

[0118] The processing module is used to perform link reconstruction processing on the processing unit link and the data unit link according to the dimensional transformation characteristics of the data unit link, so as to form the target application link.

[0119] For example, the disassembly module includes:

[0120] An extraction submodule is used to determine the application processing procedure of the initial application link and extract the processing feature information of the application processing procedure;

[0121] The disassembly submodule is used to determine the processing actions involved in the initial application link based on the processing feature information, and to disassemble the initial application link into object units based on the processing actions, and to reassemble the object units into processing unit links.

[0122] The arrangement submodule is used to generate the input matrix and output matrix corresponding to the object unit in the processing unit link according to the processing feature information, and arrange the input matrix and the output matrix according to a preset rule to form a data unit link.

[0123] For example, the arrangement submodule includes:

[0124] The first determining unit is configured to determine the data unit characteristics of the object unit in the processing unit link based on the processing feature information.

[0125] The processing unit is configured to generate an input matrix and an output matrix corresponding to the object unit based on the characteristics of the data unit, and to label the matrix dimensions of the input matrix and the output matrix.

[0126] An arrangement unit is used to arrange the input matrix and the output matrix according to a preset rule to form a data unit link based on the matrix dimension.

[0127] For example, the arrangement submodule further includes:

[0128] A marking unit is used to mark the link components in the processing unit link and the data unit link by features; wherein, the link components include the data processing unit and the data conversion unit corresponding to the input matrix and the output matrix, wherein, when marking the data conversion unit by features, the input matrix and the output matrix are marked as a whole.

[0129] For example, the processing module includes:

[0130] The comparison submodule is used to compare the matrix dimension transformation information between each adjacent data object in the processing unit link and the data unit link based on the dimension transformation characteristics of the data unit link.

[0131] The merging submodule is used to merge adjacent data objects that have a dimensional transformation relationship according to the transformation information to obtain an abstract unit;

[0132] The reconstruction submodule is used to perform link reconstruction processing on the abstract unit to form the target application link.

[0133] For example, the refactoring submodule includes:

[0134] The first reconstruction unit is used to independently mark the abstract unit, recombine and reconstruct the links where the abstract unit is located to form the first target application link;

[0135] The second reconstruction unit is used to reconstruct the abstract unit into a complete link, and to use the complete link as the second target application link.

[0136] For example, the merging submodule includes:

[0137] The second determining unit is used to determine the dimension transformation expression of the adjacent data objects based on the transformation information;

[0138] The third determining unit is used to determine that the adjacent data objects have the transformation information when there is only one value of zero in the dimension transformation expression, and to merge them to obtain the abstract unit.

[0139] The specific implementation of the application metadata processing device in this application is basically the same as the embodiments of the application metadata processing method described above, and will not be repeated here.

[0140] In addition, this application also provides an application metadata processing device. For example... Figure 4 As shown, Figure 4 This is a schematic diagram of the hardware operating environment involved in the embodiments of this application.

[0141] For example, Figure 4 This can be a schematic diagram of the hardware operating environment of the application metadata processing device.

[0142] like Figure 4 As shown, the application metadata processing device may include a processor 401, a communication interface 402, a memory 403, and a communication bus 404. The processor 401, the communication interface 402, and the memory 403 communicate with each other through the communication bus 404. The memory 403 is used to store computer programs. When the processor 401 executes the program stored in the memory 403, it implements the steps of the application metadata processing method.

[0143] The communication bus 404 mentioned in the aforementioned application metadata processing device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus 404 can be divided into an address bus, a data bus, and a control bus, etc. For ease of representation, it is shown in the figure with only one thick line, but this does not indicate that there is only one bus or one type of bus.

[0144] Communication interface 402 is used for communication between the aforementioned application metadata processing device and other devices.

[0145] The memory 403 may include random access memory (RMD) or non-volatile memory (NM), such as at least one disk storage device. Optionally, the memory 403 may also be at least one storage device located remotely from the aforementioned processor 401.

[0146] The processor 401 mentioned above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be 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, or discrete hardware components.

[0147] The specific implementation of the metadata processing device in this application is basically the same as the embodiments of the metadata processing method described above, and will not be repeated here.

[0148] Furthermore, embodiments of this application also propose a computer-readable storage medium storing an application metadata processing program, which, when executed by a processor, implements the steps of the application metadata processing method described above.

[0149] The specific implementation of the computer-readable storage medium in this application is basically the same as the embodiments of the application metadata processing method described above, and will not be repeated here.

[0150] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0151] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0152] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0153] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. An application metadata processing method, characterized by, The application metadata processing method comprises the following steps: obtaining an initial application link of a data processing execution process; extracting processing characteristic information of the initial application link, and according to the processing characteristic information, disassembling the initial application link into a processing unit link and a data unit link; wherein the processing characteristic information comprises metadata for describing a processing process of the initial application link; according to a dimension transformation characteristic of the data unit link, performing link reconstruction processing on the processing unit link and the data unit link to form a target application link; the step of according to the dimension transformation characteristic of the data unit link, performing link reconstruction processing on the processing unit link and the data unit link to form a target application link comprises: comparing transformation information of matrix dimensions between adjacent data objects in the processing unit link and the data unit link according to the dimension transformation characteristic of the data unit link; according to the transformation information, merging adjacent data objects with dimension transformation association to obtain an abstract unit; performing link reconstruction processing on the abstract unit to form a target application link.

2. The application metadata processing method of claim 1, wherein, the step of extracting processing characteristic information of the initial application link, and according to the processing characteristic information, disassembling the initial application link into a processing unit link and a data unit link comprises: determining an application processing process of the initial application link, and extracting processing characteristic information of the application processing process; according to the processing characteristic information, determining processing actions involved in the initial application link, and according to the processing actions, disassembling the initial application link into object units, and reconstructing the object units into a processing unit link; according to the processing characteristic information, generating input matrices and output matrices corresponding to object units in the processing unit link, and arranging the input matrices and the output matrices according to a preset rule to form a data unit link.

3. The application metadata processing method of claim 2, wherein, the step of according to the processing characteristic information, generating input matrices and output matrices corresponding to object units in the processing unit link, and arranging the input matrices and the output matrices according to a preset rule to form a data unit link comprises: according to the processing characteristic information, determining data unit characteristics of the object units in the processing unit link; according to the data unit characteristics, generating input matrices and output matrices corresponding to the object units, and marking matrix dimensions of the input matrices and the output matrices; according to the matrix dimensions, arranging the input matrices and the output matrices according to a preset rule to form a data unit link.

4. The application metadata processing method of claim 3, wherein, after the step of according to the data unit characteristics, generating input matrices and output matrices corresponding to the object units, and marking matrix dimensions of the input matrices and the output matrices, the method further comprises: performing feature marking on link group units in the processing unit link and the data unit link; The link forming unit comprises the data processing unit and a data conversion unit corresponding to the input matrix and the output matrix, wherein the input matrix and the output matrix are marked as a whole when the data conversion unit is marked.

5. The method of claim 1, wherein the application metadata is used to determine whether to apply the patch to the application. 5 The target application link comprises a first target application link and a second target application link, and the link reconstruction processing of the abstract unit to form the target application link comprises: The abstract unit is marked independently, and the link of the abstract unit is recombined and subjected to link reconstruction processing to form the first target application link; The abstract unit is reconstructed into a complete link, and the complete link is taken as the second target application link.

6. The application metadata processing method of claim 1, wherein, The step of merging the adjacent data objects associated with the dimension transformation according to the transformation information to obtain the abstract unit comprises: The dimension transformation expression of the adjacent data objects is determined according to the transformation information; When there is only one value of zero in the dimension transformation expression, it is determined that the adjacent data objects have the transformation information, and the adjacent data objects are merged to obtain the abstract unit.

7. An application metadata processing apparatus characterized by comprising: The application metadata processing device comprises: An acquisition module configured to acquire an initial application link of a data processing execution process; A disassembly module configured to extract processing feature information of the initial application link, and disassemble the initial application link into a processing unit link and a data unit link according to the processing feature information, wherein the processing feature information comprises metadata used to describe a processing process of the initial application link; A processing module configured to perform link reconstruction processing on the processing unit link and the data unit link according to a dimension transformation feature of the data unit link to form a target application link. The application metadata processing device is configured to: Compare transformation information of matrix dimensions between each adjacent data object in the processing unit link and the data unit link according to the dimension transformation feature of the data unit link; Merge adjacent data objects associated with the dimension transformation according to the transformation information to obtain an abstract unit; Perform link reconstruction processing on the abstract unit to form a target application link.

8. An application metadata processing apparatus characterized by comprising: The device comprises a memory, a processor, and an application metadata processing program stored on the memory and executable on the processor, and the application metadata processing program is configured to implement the steps of the application metadata processing method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an application metadata processing program, and the application metadata processing program is executed by the processor to implement the steps of the application metadata processing method according to any one of claims 1 to 6.

Images