WO2023273177A1

WO2023273177A1 - Test case processing method and apparatus, platform, and storage medium

Info

Publication number: WO2023273177A1
Application number: PCT/CN2021/136833
Authority: WO
Inventors: 史媛媛; 卢道和; 周杰; 黄涛; 陈文龙; 袁文静
Original assignee: 深圳前海微众银行股份有限公司
Priority date: 2021-06-29
Filing date: 2021-12-09
Publication date: 2023-01-05
Also published as: CN113342683B; CN113342683A

Abstract

A test case processing method and apparatus, a platform, and a storage medium. The method comprises: obtaining all front-end test cases and all front-end historical cases; calculating a feature value corresponding to a feature in each front-end test case to obtain a test feature matrix of a system version set to be tested, and calculating a feature value corresponding to a feature in each front-end historical case to obtain a historical feature matrix of each historical system version set; on the basis of a non-negative matrix factorization algorithm, performing dimension reduction processing on the test feature matrix and the historical feature matrix respectively to obtain a dimension-reduced test feature matrix and a dimension-reduced historical feature matrix; calculating a similarity between a dimension-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and a dimension-reduced feature of each front-end historical case in the dimension-reduced historical feature matrix to obtain a similarity matrix; and on the basis of the similarity matrix, determining that all the front-end test cases satisfy automated test conditions, and performing automated test on all the front-end test cases.

Description

A test case processing method, device, platform and storage medium

Cross References to Related Applications

This application is based on a Chinese patent application with application number 202110729999.3 and a filing date of June 29, 2021, and claims the priority of this Chinese patent application. The entire content of this Chinese patent application is hereby incorporated by reference into this application.

technical field

The embodiment of the present application relates to the technical field of financial technology (Fintech) data processing, and relates to but not limited to a test case processing method, a test case processing device, a test platform and a storage medium.

Background technique

With the development of computer computing, more and more technologies are applied in the financial field, and the traditional financial industry is gradually transforming into Fintech. However, due to the security and real-time requirements of the financial industry, Fintech also raises the higher requirements.

In the field of financial technology, the test cases in the financial system include front-end test cases and back-end test cases. Currently, the classification of test cases in the financial system is achieved by manual marking by testers, and the system under test in the financial system Whether the version is automatically tested depends on the experience of the tester, subjectively judging the frequency of requirement changes in the version of the system to be tested and the stability of the test case. It can be seen that the above method must rely on manual operation, at least there are problems of low efficiency and poor accuracy.

Contents of the invention

The embodiments of the present application provide a test case processing method, a test case processing device, a test platform, and a storage medium, so as to solve the problems that related technologies must rely on manual operations, and at least have low efficiency and poor accuracy.

The embodiment of this application provides a method, including:

Obtain all front-end test cases in the system version set to be tested, and all front-end historical cases in each historical system version set in at least two historical system version sets;

Calculate the eigenvalues corresponding to the features in each front-end test case in all the front-end test cases, and obtain the test feature matrix of the version set of the system to be tested;

Calculating the eigenvalues corresponding to the features in each of the front-end historical cases in all the front-end historical cases in the set of each historical system version, to obtain the historical feature matrix of each set of historical system versions;

Based on a non-negative matrix factorization algorithm, performing dimensionality reduction processing on the test feature matrix to obtain a dimensionally reduced test feature matrix;

Based on the non-negative matrix decomposition algorithm, performing dimensionality reduction processing on the historical feature matrix to obtain a dimensionally reduced historical feature matrix;

Calculating the similarity between the dimension-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced feature of each front-end historical case in the dimension-reduced historical feature matrix, to obtain A similarity matrix; wherein, the n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the version set of the system to be tested;

Based on the similarity matrix, when it is determined that all the front-end test cases meet the automated test conditions, all the front-end test cases in the version set of the system under test are automatically tested.

An embodiment of the present application provides a device, including:

An acquisition module, configured to acquire all front-end test cases in the system version set to be tested, and all front-end historical cases in each historical system version set in at least two historical system version sets;

A processing module, configured to calculate the eigenvalues corresponding to the features in each of the front-end test cases in all the front-end test cases, and obtain the test feature matrix of the version set of the system to be tested;

The processing module is also used to calculate the feature value corresponding to the feature in each front-end historical case in all the front-end historical cases in the set of each historical system version, and obtain the historical feature of each historical system version set matrix;

The processing module is also used to perform dimensionality reduction processing on the test feature matrix based on a non-negative matrix factorization algorithm to obtain a dimensionality-reduced test feature matrix;

The processing module is further configured to perform dimensionality reduction processing on the historical feature matrix based on the non-negative matrix factorization algorithm, to obtain a dimensionally reduced historical feature matrix;

The processing module is also used to calculate the dimensionality-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and the dimensionality reduction of each front-end historical case in the dimension-reduced historical feature matrix The degree of similarity of the features after is obtained similarity matrix; Wherein, said n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the described system version set to be tested;

The processing module is further configured to perform automated testing on all front-end test cases in the version set of the system under test when it is determined that all front-end test cases meet the automated test conditions based on the similarity matrix.

The embodiment of the present application provides a test platform, including:

memory for storing executable instructions;

When the processor is configured to execute the executable instructions stored in the memory, the above method is implemented.

An embodiment of the present application provides a storage medium, which stores executable instructions, and is used to cause a processor to implement the above method when executed.

The embodiment of the present application has the following beneficial effects:

After obtaining all front-end test cases and all front-end historical cases in the version set of the system to be tested, the test platform of this application first calculates the characteristic value corresponding to the feature in each front-end test case, and obtains the test features of the version set of the system to be tested Matrix, calculate the eigenvalues corresponding to the features in each front-end historical case, and obtain the historical feature matrix of each historical system version set; secondly, use the non-negative matrix algorithm to perform dimension reduction processing on the test feature matrix and the historical feature matrix respectively, And perform similarity processing on the test feature matrix after dimensionality reduction and the historical feature matrix after dimensionality reduction to obtain a similarity matrix, and then according to the similarity matrix, when all front-end test cases meet the automated test conditions, all front-end test cases Run automated tests. In this way, this application solves the problem that related technologies must rely on manual operation and uncertainty brought by manual subjectivity, and at least has the problems of low efficiency and poor accuracy; realizes the establishment of a unified standard for the automatic execution of front-end test cases, and improves The accuracy of judgment is improved, and at the same time, it does not need to rely on manual operation, which improves the processing efficiency.

Description of drawings

Fig. 1 is a schematic diagram of an optional architecture of the test platform provided by the embodiment of the present application;

Fig. 2 is an optional schematic flow chart of the test case processing method provided by the embodiment of the present application;

Fig. 3 is an optional schematic flow chart of the test case processing method provided by the embodiment of the present application;

Fig. 4 is an optional flowchart for training a classifier in a front-end case provided by an embodiment of the present application;

Fig. 5 is an optional schematic flow chart of the test case processing method provided by the embodiment of the present application;

FIG. 6 is an optional schematic flow chart of a test case processing method provided in an embodiment of the present application;

Fig. 7 is an optional schematic flow chart of the test case processing method provided by the embodiment of the present application;

Fig. 8 is a schematic flowchart of an optional test case processing method provided by the embodiment of the present application.

detailed description

In order to make the purpose, technical solutions and advantages of the application clearer, the application will be further described in detail below in conjunction with the accompanying drawings. All other embodiments obtained under the premise of creative labor belong to the scope of protection of this application.

In the following description, references to "some embodiments" describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or a different subset of all possible embodiments, and Can be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meaning as commonly understood by those skilled in the technical field of the embodiments of the present application. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

The following describes exemplary applications provided by the embodiments of the present application, which may be implemented as servers. In the following, an exemplary application of the implementation will be described.

Referring to FIG. 1, FIG. 1 is a schematic structural diagram of a test platform 100 provided by an embodiment of the present application. The test platform 100 shown in FIG. Various components in the test platform 100 are coupled together through the bus system 140 . It can be understood that the bus system 140 is used to realize connection and communication between these components. In addition to the data bus, the bus system 140 also includes a power bus, a control bus and a status signal bus. However, for clarity of illustration, the various buses are labeled as bus system 140 in FIG.

Processor 110 can be a kind of integrated circuit chip, has signal processing capability, such as general-purpose processor, digital signal processor (DSP, Digital Signal Processor), or other programmable logic device, discrete gate or transistor logic device, discrete hardware Components, etc., wherein the general-purpose processor can be a microprocessor or any conventional processor, etc.

User interface 130 includes one or more output devices 131 that enable presentation of media content, including one or more speakers and/or one or more visual displays. The user interface 130 also includes one or more input devices 132, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

Memory 150 may be removable, non-removable or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, and the like. Memory 150 optionally includes one or more storage devices located physically remote from processor 110 . Memory 150 includes volatile memory or nonvolatile memory, and may include both volatile and nonvolatile memory. The non-volatile memory can be read-only memory (Read Only Memory, ROM), and the volatile memory can be random access memory (Random Access Memory, RAM). The memory 150 described in the embodiment of the present application is intended to include any suitable type of memory. In some embodiments, the memory 150 is capable of storing data to support various operations, examples of which include programs, modules, and data structures, or subsets or supersets thereof, as exemplified below.

Operating system 151, including system programs for processing various basic system services and performing hardware-related tasks, such as framework layer, core library layer, driver layer, etc., for implementing various basic services and processing hardware-based tasks;

The network communication module 152 is used to reach other computing devices via one or more (wired or wireless) network interfaces 120. Exemplary network interfaces 120 include: Bluetooth, Wireless Compatibility Authentication (WiFi), and Universal Serial Bus ( Universal Serial Bus, USB), etc.;

The input processing module 153 is configured to detect one or more user inputs or interactions from one or more of the input devices 132 and translate the detected inputs or interactions.

In some embodiments, the device provided by the embodiment of the present application can be realized by software. FIG. 1 shows a test case processing device 154 stored in the memory 150. The test case processing device 154 can be the The test case processing device, which can be software in the form of programs and plug-ins, includes the following software modules: acquisition module 1541, processing module 1542, these modules are logical, so any combination or further split. The function of each module will be explained below.

In other embodiments, the device provided in the embodiment of the present application may be implemented in hardware. As an example, the device provided in the embodiment of the present application may be a processor in the form of a hardware decoding processor, which is programmed to execute the In the test case processing method provided by the embodiment, for example, the processor in the form of a hardware decoding processor can adopt one or more Application Specific Integrated Circuits (Application Specific Integrated Circuit, ASIC), DSP, Programmable Logic Device (Programmable Logic Device, PLD), Complex Programmable Logic Device (Complex Programmable Logic Device, CPLD), Field Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other electronic components.

Before describing the test case processing method provided in the embodiment of the present application, the application background and related technologies implemented in the present application are briefly introduced.

With the rise of the Internet industry and the Internet of Things, the software development of the application layer software in the test platform, such as the application system of the test platform and the application program (Application, APP) of the test platform, is not released at one time, but is iterative. Continuous incremental development and incremental release. The application layer software of the new version system is the application layer software in the system under test, which is obtained by modifying a certain module in the historical version system and/or adding a certain module, which makes the web front end of the system under test often The changes are relatively large, but the logic of some modules in the reused historical version system has not changed, that is, the back-end implementation has not changed. Here, at present, the classification of the front-end and back-end in the system under test is mainly realized through manual marking by testers, and the functional verification of the web front-end of the system under test must be tested through automated scripts. Whether the functional verification of the web front-end of the test system is automatically tested mainly depends on the experience of the testers, subjectively judging the frequency of requirements changes in the version of the system under test and the stability of the test cases. It can be seen that at least there are problems of low efficiency and poor accuracy in the above method.

The test case processing method provided in the embodiment of the present application will be described below in conjunction with the exemplary application and implementation of the test platform 100 provided in the embodiment of the present application. Referring to FIG. 2, FIG. 2 is an optional flow chart of the test case processing method provided in the embodiment of the present application, which will be described in conjunction with the steps shown in FIG. 2,

Step 201. Acquire all front-end test cases in the system version set to be tested, and all front-end historical cases in each of the at least two historical system version sets.

In this embodiment of the application, the system to be tested is a system that tests the cases contained in the system before release, and the historical system is a system that has been successfully released. The system version to be tested can be understood as the system corresponding to the current version number, and the historical system version can be understood as the system corresponding to the version numbers of different historical periods, and there can be multiple historical system versions. In an achievable scenario, developers iteratively and incrementally develop the legacy system to obtain the system under test. Exemplarily, the developer obtains the system under test by modifying a certain functional module in the historical system and/or adding a certain functional module.

In the embodiment of the present application, the front-end test cases are all test cases included in the web front-end of the system to be tested, and all the test cases included in the web front-end of the system to be tested form the version set of the system to be tested; All historical cases included in the web front end of the system, and all historical cases included in the web front end of each historical system constitute each historical system version set. Here, a case can be understood as a functional module of the system front-end page. Exemplarily, a case can be a registration module of the system front-end page, and a case can also be a login module of the system front-end page. Here, the total number of front-end test cases in the system version set to be tested may be the same as or different from the total number of front-end historical cases in each of the at least two historical system version sets. In the embodiment of the present application, the total number of front-end test cases in the system version set to be tested is the same as the total number of front-end historical cases in each of the at least two historical system version sets.

In this embodiment of the application, the test platform obtains all the front-end test cases in the system version set to be tested. Further, the test platform also obtains at least two historical system version sets corresponding to the system to be tested. In each historical system version set All front-end history cases for .

Step 202. Calculate the eigenvalues corresponding to the features in each front-end test case in all the front-end test cases, and obtain the test feature matrix of the system version set to be tested.

In the embodiment of the present application, the feature value corresponding to a feature in each front-end test case in all front-end test cases is used to represent the importance of the feature in the front-end test case.

In the embodiment of the present application, the test platform can calculate the eigenvalues corresponding to the features in each front-end test case in all front-end test cases based on the word frequency-reverse file frequency algorithm, and obtain the test feature matrix of the system version set to be tested.

Here, the term frequency-inverse document frequency (TF-IDF) algorithm is a statistical analysis method for words, which is used to evaluate the importance of a word to a document set or a corpus. The importance of a word is directly proportional to the number of times it appears in the article, and inversely proportional to the number of times it appears in the corpus. TF=(the number of times a word appears in the document/the total number of words in the document), IDF=log(the total number of documents in the corpus/the number of documents containing the word+1), the result of TF-IDF is TF×IDF.

In the embodiment of this application, the test platform is based on the TF-IDF algorithm to calculate the feature value corresponding to the _mth feature in the nth case of all front-end test cases N, and then all the features in all front-end test cases correspond to The eigenvalues of , generate the test feature matrix of the system version set under test. Here, the test feature matrix

Among them, the matrix size of the test feature matrix X of the system version set to be tested is N×M, n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the system version set to be tested; m is A positive integer greater than or equal to 1 and less than or equal to N, where M is the number of features of each front-end test case.

Step 203: Calculate the eigenvalues corresponding to the features in each front-end historical case in all front-end historical cases in each historical system version set, and obtain the historical feature matrix of each historical system version set.

In the embodiment of the present application, the feature value corresponding to the feature in each front-end historical case in all front-end historical cases in each historical system version set is used to represent the importance of the feature in the front-end historical case. Here, the total number of features in each front-end test case in all front-end test cases may be the same as or different from the total number of features in each front-end historical case in each historical system version set. In the embodiment of this application, the total number of features in each front-end test case in all front-end test cases is the same as the total number of features in each front-end historical case in each historical system version set as an example. .

In the embodiment of this application, the test platform is based on the TF-IDF algorithm, and calculates the m-th feature of the n-th front-end historical case in all front-end historical cases N in the k-th historical system version set in at least two historical system version sets corresponding eigenvalues

Then, from the eigenvalues corresponding to all the features in all front-end historical cases in the kth historical system version set, a historical feature matrix of the kth historical system version set is generated. Here, the historical feature matrix

Among them, the matrix size of the historical feature matrix P _k corresponding to the kth historical system version set is M×N, n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the front-end historical case in each historical system version set m is a positive integer greater than or equal to 1 and less than or equal to N, and M is the number of features of each front-end historical case; k is a positive integer greater than or equal to 1 and less than or equal to K, and K is the total number of all historical system version sets.

It should be noted that step 202 and step 203 may be performed simultaneously, or step 202 and step 203 may be performed sequentially, which is not specifically limited in this application.

Step 204 , based on the non-negative matrix factorization algorithm, perform dimensionality reduction processing on the test feature matrix to obtain a dimensionality-reduced test feature matrix.

In the embodiment of the present application, the test platform calculates the eigenvalues corresponding to the features in each front-end test case in all front-end test cases based on the word frequency-reverse file frequency algorithm, and obtains the test feature matrix of the system version set to be tested. In order to efficiently process the data stored in the test feature matrix, the test platform performs dimensionality reduction processing on the test feature matrix through the non-negative matrix decomposition algorithm, and obtains the test feature matrix after dimensionality reduction. In this way, in the case of ensuring that each element in the dimension-reduced test feature matrix is a non-negative value, the dimension-reduced test feature matrix is used to replace the original test feature matrix. At this time, the dimension-reduced test feature matrix The feature matrix is processed, which not only reduces the storage space, but also reduces the calculation amount of computer resources.

Here, the non-negative matrix factorization (Non-negative matrix factorization, NMF) algorithm is a matrix factorization method under the constraint that all elements in the matrix are non-negative numbers, that is, for a given non-negative matrix A, the NMF algorithm can find a A non-negative matrix U and a non-negative matrix V with smaller dimensions, that is, A≈U×V, to obtain a dimensionally reduced non-negative matrix U. It should be noted that, in mathematics, from a calculation point of view, it is correct to have negative values in the decomposition results, but negative value elements are often meaningless in practical problems. For example, in the embodiment of the present application, the eigenvalues corresponding to the features in each case cannot have negative features. Therefore, using the NMF algorithm can enable the test platform to process according to actual problems.

Step 205 , based on the non-negative matrix factorization algorithm, perform dimensionality reduction processing on the historical feature matrix to obtain a dimensionally reduced historical feature matrix.

In the embodiment of this application, the test platform calculates the feature value corresponding to the feature in each front-end historical case in all front-end historical cases in each historical system version set based on the word frequency-reverse document frequency algorithm, and obtains each historical system version set In the case of the historical feature matrix, in order to efficiently process the data stored through the historical feature matrix, the test platform uses the non-negative matrix factorization algorithm to reduce the dimensionality of the historical feature matrix to obtain the historical feature matrix after dimensionality reduction. In this way, in the case of ensuring that each element in the historical feature matrix after dimensionality reduction is a non-negative value, the historical feature matrix after dimensionality reduction is used to replace the original historical feature matrix. At this time, the historical feature matrix after dimensionality reduction The feature matrix is processed, which not only reduces the storage space, but also reduces the calculation amount of computer resources.

It should be noted that step 204 and step 205 may be performed simultaneously, and step 204 and step 205 may also be performed sequentially, which is not specifically limited in this application.

Step 206, calculate the similarity between the dimensionality-reduced features of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced features of each front-end historical case in the dimension-reduced historical feature matrix, and obtain similarity degree matrix.

Wherein, n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the system version set to be tested.

In the embodiment of the present application, the test platform performs dimensionality reduction processing on the test feature matrix based on the non-negative matrix decomposition algorithm, obtains the dimensionality-reduced test feature matrix, and performs dimensionality reduction processing on the historical feature matrix based on the non-negative matrix decomposition algorithm , when the dimensionality-reduced historical feature matrix is obtained, calculate the dimensionality-reduced feature of the nth front-end test case in the dimension-reduced test feature and the dimensionality reduction of each front-end historical case in the dimensionality-reduced historical feature matrix The similarity of the final features is obtained to obtain a similarity matrix, so that the test platform can judge whether all the front-end test cases in the system version set under test meet the automated test conditions based on the similarity matrix.

Step 207 , based on the similarity matrix, when it is determined that all front-end test cases meet the automated test conditions, perform automated tests on all front-end test cases in the system version set to be tested.

In the embodiment of the present application, the automated test condition is the condition that the front-end test case in the system version set to be tested can use the automated script to perform the automated test.

In the embodiment of the present application, the test platform calculates the dimensionality-reduced features of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced features of each front-end historical case in the dimension-reduced historical feature matrix In the case of obtaining the similarity matrix, based on the similarity matrix, when it is determined that all front-end test cases meet the automated test conditions, all front-end test cases in the system version set to be tested are automatically tested through automated scripts.

In the test case processing method provided by the embodiment of the present application, after the test platform obtains all front-end test cases and all front-end historical cases in the system version set to be tested, first, calculates the feature value corresponding to the feature in each front-end test case, Obtain the test feature matrix of the system version set to be tested, calculate the eigenvalues corresponding to the features in each front-end historical case, and obtain the historical feature matrix of each historical system version set; secondly, through the non-negative matrix algorithm, the test feature matrix and Dimensionality reduction processing is performed on the historical feature matrix, and the similarity processing is performed on the reduced test feature matrix and the reduced historical feature matrix to obtain a similarity matrix, and then according to the similarity matrix, it is determined that all front-end test cases meet the requirements of automation When testing conditions, automate tests for all front-end test cases. In this way, this application solves the problem that related technologies must rely on manual operation and uncertainty brought by manual subjectivity, and at least has the problems of low efficiency and poor accuracy; realizes the establishment of a unified standard for the automatic execution of front-end test cases, and improves The accuracy of judgment is improved, and at the same time, it does not need to rely on manual operation, which improves the processing efficiency.

Referring to FIG. 3, FIG. 3 is an optional schematic flowchart of the test case processing method provided in the embodiment of the present application, which will be described in conjunction with the steps shown in FIG. 3,

Step 301. Obtain all test cases in the version set of the system to be tested.

Step 302: Input all the test cases in the version set of the system under test into the trained classifier, and obtain the front-end test cases in the version set of the system under test output by the trained classifier.

In the embodiment of the present application, all test cases in the system version set to be tested include front-end test cases and back-end test cases.

Here, the trained classifier is used to classify all the test cases in the version set of the system under test to obtain the front-end test cases among all the test cases in the version set of the system under test. A classifier is a method of data mining. The classifier can map the data records in the database to one of the given categories, so that it can be applied to data prediction. Here, a classifier is a general term for methods for classifying samples in data mining. Classification methods include algorithms such as decision trees, logistic regression, naive Bayesian, and neural networks. Classification methods also include support vector machines (support vector machines, SVM) algorithm.

In an achievable scenario, as shown in Figure 4, the test platform obtains all marked front-end historical cases and back-end historical cases in each historical system version set in at least two historical system version sets as a training sample set Afterwards, the test platform uses the TF-IDF method to calculate the eigenvalues corresponding to each feature in the front-end historical cases in each historical system version set in at least two historical system version sets, and the corresponding eigenvalues for each feature in the back-end historical cases eigenvalues, so as to obtain the training sample feature matrix. The test platform trains the classifier through the training sample feature matrix to obtain a trained classifier.

In a realizable application scenario, SVM is used as an example to illustrate. The test platform uses SVM for classification training. SVM is a binary classification model, which mainly looks for the classifier with the largest interval in the feature space, combined with the kernel function , can classify nonlinear features, and the actual implementation can be converted into a problem of solving convex quadratic programming. First, construct a convex quadratic programming problem:

The constraints are:

Combined with the sequential minimal optimization algorithm (Sequential minimal optimization, SMO), the Lagrangian parameters are obtained

Optimal solution

Again, according to the KKT condition

and constraints

The optimal solution of the Lagrangian parameters can be obtained

corresponding

Take the classification decision function as:

Here, b is the parameter to confirm the classification decision function corresponding to each sample, and the exponential kernel function

Here, δ is the distance between features, and l is the hyperparameter of the kernel function. The SVM classifier corresponding to the exponential kernel function is a binary classifier of the exponential function, then the final classification decision function is

Based on this, the sample set is trained to obtain the final classifier.

In the embodiment of the present application, as shown in FIG. 4, after the test platform obtains all the test cases in the system version set to be tested, the test platform uses the TF-IDF method to calculate each test case in all the test cases in the system version set to be tested. The eigenvalue corresponding to a feature is obtained to obtain the test feature matrix, and the test feature matrix is input into the trained classifier to obtain the front-end test cases in the version set of the system under test output by the trained classifier.

In some embodiments, as shown in FIG. 4, the test platform inputs all test cases in the version set of the system under test into the trained classifier, and obtains the front end in the version set of the system under test output by the trained classifier. After the test cases, pass the obtained front-end test cases through the jieba model, traverse each front-end test case, and mark the traversed front-end test cases as "front-end". In this way, through machine learning and classifiers, front-end and back-end test cases are automatically classified, freeing manpower, saving labor costs, reducing the risk of misclassification caused by manual marking, and improving processing efficiency.

In some embodiments, as shown in FIG. 4, after the test platform marks the front-end test cases and the back-end test cases in the version set of the system to be tested, in order to improve the classification accuracy of the classifier, the test platform will use the version of the system to be tested The marked front-end test cases and back-end test cases in the collection are re-added to the training sample set, and the classifier is continuously trained, so that the classifier can classify the front-end test cases more accurately and quickly.

Step 303. Obtain all front-end historical cases in each historical system version set in at least two historical system version sets.

Step 304: Calculate the eigenvalues corresponding to the features in each front-end test case in all the front-end test cases, and obtain the test feature matrix of the system version set to be tested.

Step 305: Calculate the eigenvalues corresponding to the features in each front-end historical case in all front-end historical cases in each historical system version set, and obtain the historical feature matrix of each historical system version set.

Step 306 , based on the non-negative matrix factorization algorithm, perform dimensionality reduction processing on the test feature matrix to obtain a dimensionality-reduced test feature matrix.

In some embodiments, step 306 is based on the non-negative matrix factorization algorithm to perform dimensionality reduction processing on the test feature matrix to obtain the dimensionality-reduced test feature matrix, which can be performed through steps 401 to 403 shown in FIG. 5; or steps 401 to 403. Step 402 and step 404 to step 407; or step 401 to step 402, step 404 to step 406 and step 408 to step 411 to achieve:

Step 401 : Decompose the test feature matrix through a non-negative matrix factorization algorithm based on the determined feature numbers after dimension reduction of the test feature matrix to obtain a test projection matrix and a test fundamental matrix.

In the embodiment of this application, the value range of the test platform based on the feature number S after dimensionality reduction is

Select a positive integer S to determine the feature number S after dimension reduction, and decompose the test feature matrix X through the non-negative matrix decomposition algorithm, and randomly generate the test projection matrix W1 and the test basic matrix B1. Here, the size of the test projection matrix W1 is N×S, and the size of the test fundamental matrix B1 is S×M.

Step 402. Obtain a first product matrix obtained by multiplying the test projection matrix and the test fundamental matrix.

Step 403: If the first difference matrix obtained by subtracting the first product matrix from the test feature matrix conforms to the difference threshold matrix, determine the test projection matrix corresponding to the first difference matrix as the reduced-dimensional test feature matrix.

In the embodiment of the present application, the difference threshold matrix is used to determine the difference matrix between the test feature matrix and the first product matrix obtained by multiplying the test projection matrix and the test fundamental matrix. Exemplarily, the value of each element in the difference threshold matrix may be 10 ^-6 .

In the embodiment of the present application, the test platform decomposes the test feature matrix X through the non-negative matrix decomposition algorithm based on the determined feature number S after dimensionality reduction of the test feature matrix, and obtains the test projection matrix W1 and the test basic matrix B1 , to obtain the first product matrix Q1 obtained by multiplying the test projection matrix W1 and the test fundamental matrix B1. If the first difference matrix E1 obtained by subtracting the first product matrix Q1 from the test feature matrix X conforms to the difference threshold matrix E, since the test platform needs to find the test projection matrix that satisfies the difference threshold matrix E, it is the test feature after dimensionality reduction At this time, because the first difference matrix E1 satisfies the condition, the test projection matrix W1 corresponding to the first difference matrix E1 is the test feature matrix after dimensionality reduction

Here, the test feature matrix after dimensionality reduction

Among them, the test feature matrix after dimensionality reduction

The size of is N×S.

Step 404: If the first difference matrix obtained by subtracting the first product matrix from the test feature matrix does not conform to the difference threshold matrix, adjust each element in the test projection matrix through the projection matrix adjustment model to obtain the adjusted test projection matrix.

In the embodiment of this application, the projection matrix adjustment model is:

Among them, W' _ns is the element in row n and column s in the adjusted test projection matrix, W is the test projection matrix, W _ns is the element in row n and column s in the test projection matrix, X is the test feature matrix, B is the test fundamental matrix, B ^T is the transpose matrix of the test fundamental matrix, (XB ^T ) _ns is the nth row s of the matrix obtained by multiplying the test feature matrix X and the transpose matrix B ^T of the test fundamental matrix The elements of the column, (WBB ^T ) _ns are the elements of the nth row and the sth column in the matrix obtained by multiplying the test projection matrix W, the test fundamental matrix B and the transpose matrix B ^T of the test fundamental matrix.

In other embodiments of the present application, the test platform determines that the first difference matrix E1 obtained by subtracting the first product matrix Q1 from the test feature matrix X does not meet the difference threshold matrix E, and adjusts the model for each of the test projection matrix W1 through the projection matrix. One element is adjusted to obtain the adjusted test projection matrix W11.

Step 405: Adjust each element in the test fundamental matrix through the fundamental matrix adjustment model to obtain an adjusted test fundamental matrix.

In the embodiment of this application, the basic matrix adjustment model is:

Among them, B'sm is the element of the sth row and the mth column in the adjusted test fundamental matrix, B is the test fundamental matrix, B _sm is the element of the _sth row and the mth column in the test fundamental matrix, X is the test feature matrix, W is the test projection matrix, W ^T is the transpose matrix of the test projection matrix, (W ^T X) _sm is the matrix obtained by multiplying the transpose matrix W ^T of the test projection matrix by the test feature matrix X The elements in column m, (W ^T WB ) _sm are the elements in row s and column m in the matrix obtained by multiplying the transposition matrix W ^T of the test projection matrix, the test projection matrix W and the test fundamental matrix B.

In the embodiment of the present application, the test platform adjusts each element in the basic projection matrix B1 through the basic matrix adjustment model to obtain the adjusted test basic matrix B11.

Step 406: Obtain a second product matrix obtained by multiplying the adjusted test projection matrix and the adjusted test fundamental matrix.

Step 407: If the second difference matrix obtained by subtracting the second product matrix from the test feature matrix conforms to the difference threshold matrix, determine that the adjusted test projection matrix corresponding to the second difference matrix is the reduced-dimensional test feature matrix.

In the embodiment of the present application, the test platform obtains the second product matrix Q2 obtained by multiplying the adjusted test projection matrix W11 and the adjusted test fundamental matrix B11; The difference matrix E2 conforms to the difference threshold matrix E. Since the test platform needs to find the test projection matrix corresponding to the difference threshold matrix E, it is the test feature matrix after dimensionality reduction. At this time, because the second difference matrix E2 satisfies the condition, then Determine the adjusted test projection matrix W11 corresponding to the second difference matrix E2 as the test feature matrix after dimensionality reduction

Step 408: If the second difference matrix obtained by subtracting the second product matrix from the test feature matrix does not conform to the difference threshold matrix, adjust each element in the adjusted test projection matrix through the projection matrix adjustment model to obtain a new adjustment After the test projection matrix.

Step 409: Adjust each element in the adjusted test fundamental matrix through the fundamental matrix adjustment model to obtain a new adjusted test fundamental matrix.

Step 410: Obtain a third product matrix obtained by multiplying the newly adjusted test projection matrix and the newly adjusted test fundamental matrix.

Step 411: If the third difference matrix obtained by subtracting the third product matrix from the test feature matrix conforms to the difference threshold matrix, determine that the newly adjusted test projection matrix corresponding to the third difference matrix is the reduced-dimensional test feature matrix.

In the embodiment of the present application, first, the test platform determines that the second difference matrix E2 obtained by subtracting the second product matrix Q2 from the test feature matrix X does not meet the difference threshold matrix E, and the adjusted test projection matrix is adjusted by the projection matrix adjustment model Each element in W11 is adjusted to obtain the newly adjusted test projection matrix W12; secondly, the test platform adjusts each element in the adjusted test basic matrix B11 through the basic matrix adjustment model to obtain the newly adjusted test The basic matrix B12; then, the test platform obtains the third product matrix Q3 obtained by multiplying the newly adjusted test projection matrix W12 and the newly adjusted test basic matrix B12; finally, the test platform determines the test feature matrix X minus the third product The third difference matrix E3 obtained by matrix Q3 conforms to the difference threshold matrix E, and the newly adjusted test projection matrix W12 corresponding to the third difference matrix E3 is determined to be the test feature matrix after dimensionality reduction

It should be noted that, in the embodiment of the present application, for the dimensionality-reduced test feature matrix

The determination of , can be that the test platform adjusts the test projection matrix once to obtain the dimensionality-reduced test feature matrix that satisfies the conditions, or it can be that the test platform obtains the dimensionality-reduced feature matrix that satisfies the conditions after multiple adjustments to the test projection matrix Test feature matrix. That is to say, the present application does not specifically limit the number of adjustment cycles, and the dimensionality-reduced test feature matrix that satisfies the conditions shall prevail. Here, because the test platform needs to find the test projection matrix corresponding to the difference threshold matrix E as the test feature matrix after dimensionality reduction, at this time, because the third difference matrix E3 satisfies the condition, then determine the third difference matrix E3 corresponding to The newly adjusted test projection matrix W12 is the test feature matrix after dimensionality reduction

In the embodiment of the present application, the test platform replaces the original test feature matrix with the reduced-dimensional test feature matrix, which not only reduces the storage space, but also reduces the calculation amount and computational complexity of computer resources, and the feature representation power is improved. At the same time Whether automated testing is performed on all front-end test cases, and accurate data is provided as the basis for calculation.

Step 307 , based on the non-negative matrix factorization algorithm, perform dimensionality reduction processing on the historical feature matrix to obtain a dimensionally reduced historical feature matrix.

In some embodiments, step 307 is based on the non-negative matrix factorization algorithm to perform dimensionality reduction processing on the historical feature matrix to obtain the dimensionality-reduced historical feature matrix, which can be performed through steps 501 to 503 shown in FIG. 6; or steps 501 to 503. Step 502 and step 504 to step 507; or step 501 to step 502, step 504 to step 506 and step 508 to step 511 to achieve:

Step 501 : Based on the determined feature numbers of the historical feature matrix after dimension reduction, the historical feature matrix is decomposed through a non-negative matrix factorization algorithm to obtain a historical projection matrix and a historical fundamental matrix.

Select a positive integer S to determine the feature number S after dimension reduction, and decompose the historical feature matrix P _k through the non-negative matrix decomposition algorithm, and randomly generate the historical projection matrix W2 and historical basic matrix B2. Wherein, the size of the historical projection matrix W2 is N×S, and the size of the historical basic matrix B2 is S×M.

Step 502. Obtain a fourth product matrix obtained by multiplying the historical projection matrix and the historical fundamental matrix.

Step 503 : If the fourth difference matrix obtained by subtracting the fourth product matrix from the historical feature matrix conforms to the difference threshold matrix, determine that the historical projection matrix corresponding to the fourth difference matrix is the historical feature matrix after dimensionality reduction.

In the embodiment of the present application, the difference threshold matrix is used to determine the difference matrix between the historical feature matrix and the first product matrix obtained by multiplying the historical projection matrix and the historical fundamental matrix. Exemplarily, the value of each element in the difference threshold matrix may be 10 ^-6 .

In the embodiment of this application, the test platform decomposes the historical feature matrix P _k through the non-negative matrix decomposition algorithm based on the feature number S after the dimensionality reduction of the determined historical feature matrix, and obtains the situation of the historical projection matrix W2 and the historical basic matrix B2 Next, obtain the fourth product matrix Q4 obtained by multiplying the historical projection matrix W2 and the historical fundamental matrix B2. If the fourth difference matrix E4 obtained by subtracting the fourth product matrix Q4 from the historical feature matrix _Pk conforms to the difference threshold matrix E, since the test platform needs to find the historical projection matrix that satisfies the difference threshold matrix E, it is the dimension-reduced history Feature matrix, at this time, because the fourth difference matrix E4 satisfies the condition, the historical projection matrix W2 corresponding to the fourth difference matrix E4 is the historical feature matrix after dimensionality reduction

Here, the historical feature matrix after dimensionality reduction

Among them, the historical feature matrix after dimensionality reduction

The size is S×N.

Step 504: If the fourth difference matrix obtained by subtracting the fourth product matrix from the historical feature matrix does not conform to the difference threshold matrix, adjust each element in the historical projection matrix through the projection matrix adjustment model to obtain the adjusted historical projection matrix.

Step 505: Adjust each element in the historical basic matrix through the basic matrix adjustment model to obtain an adjusted historical basic matrix.

In the embodiment of the present application, the test platform determines that the fourth difference matrix E4 obtained by subtracting the fourth product matrix Q4 from the historical feature matrix _Pk does not meet the difference threshold matrix E, and adjusts the model for each historical projection matrix W2 through the projection matrix. One element is adjusted to obtain the adjusted historical projection matrix W21. Further, each element in the basic projection matrix B2 is adjusted through the basic matrix adjustment model to obtain the adjusted basic projection matrix B21.

Step 506: Obtain a fifth product matrix obtained by multiplying the adjusted historical projection matrix and the adjusted historical fundamental matrix.

Step 507: If the fifth difference matrix obtained by subtracting the fifth product matrix from the historical feature matrix conforms to the difference threshold matrix, determine that the adjusted historical projection matrix corresponding to the fifth difference matrix is the reduced dimensionality historical feature matrix.

In the embodiment of the present application, the test platform obtains the fifth product matrix Q5 obtained by multiplying the adjusted historical projection _matrix W21 and the adjusted historical fundamental matrix B21; The five-difference matrix E5 conforms to the difference threshold matrix E. Since the test platform needs to find that the historical projection matrix corresponding to the difference threshold matrix E is the historical feature matrix after dimensionality reduction, at this time, because the fifth difference matrix E5 satisfies the condition, Then determine that the adjusted historical projection matrix W21 corresponding to the fifth difference matrix E5 is the historical feature matrix after dimensionality reduction

Step 508: If the fifth difference matrix obtained by subtracting the fifth product matrix from the historical feature matrix does not conform to the difference threshold matrix, adjust each element in the adjusted historical projection matrix through the projection matrix adjustment model to obtain a new adjusted The post-historical projection matrix.

Step 509: Adjust each element in the adjusted historical basic matrix through the basic matrix adjustment model to obtain a new adjusted historical basic matrix.

Step 510: Obtain a sixth product matrix obtained by multiplying the newly adjusted historical projection matrix and the newly adjusted historical fundamental matrix.

Step 511 : If the sixth difference matrix obtained by subtracting the sixth product matrix from the historical feature matrix conforms to the difference threshold matrix, determine that the newly adjusted historical projection matrix corresponding to the sixth difference matrix is the dimensionality-reduced historical feature matrix.

In the embodiment of the present application, first, the test platform determines that the fifth difference matrix E5 obtained by subtracting the fifth product matrix Q5 from the historical feature matrix _Pk does not conform to the difference threshold matrix E, and adjusts the model through the projection matrix to adjust the historical projection Each element in the matrix W21 is adjusted to obtain the newly adjusted historical projection matrix W22; secondly, the test platform adjusts each element in the adjusted historical basic matrix B21 through the basic matrix adjustment model to obtain the newly adjusted The historical basic matrix B22; then, the test platform obtains the sixth product matrix Q6 obtained by multiplying the newly adjusted historical projection matrix W22 and the newly adjusted historical basic matrix B22; finally, the test platform determines the historical feature matrix P _k minus the first The sixth difference matrix E6 obtained by the six product matrix Q6 conforms to the difference threshold matrix E, and the newly adjusted historical projection matrix W22 corresponding to the sixth difference matrix E6 is determined to be the historical feature matrix after dimensionality reduction

It should be noted that, in the embodiment of the present application, for the dimensionality-reduced historical feature matrix

The determination of , can be the dimensionality-reduced historical feature matrix that satisfies the conditions after the test platform adjusts the historical projection matrix once, or the dimensionality-reduced feature matrix that meets the conditions after the test platform adjusts the historical projection matrix many times. Historical feature matrix. That is to say, the present application does not specifically limit the number of adjustment cycles, and the dimensionality-reduced historical feature matrix that satisfies the conditions is obtained. Here, since the test platform needs to find that the historical projection matrix corresponding to the difference threshold matrix E is the historical feature matrix after dimensionality reduction, at this time, because the sixth difference matrix E6 satisfies the condition, it is determined that the sixth difference matrix E6 corresponds to The newly adjusted historical projection matrix W22 is the historical feature matrix after dimensionality reduction

In the embodiment of this application, the test platform replaces the original historical feature matrix with the historical feature matrix after dimensionality reduction, which not only reduces the storage space, but also reduces the calculation amount and computational complexity of computer resources, and the feature representation power is improved. At the same time Whether automated testing is performed on all front-end test cases, and accurate data is provided as the basis for calculation.

Step 308, calculating the similarity between the dimensionality-reduced features of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced features of each front-end historical case in the dimension-reduced historical feature matrix, to obtain the similarity degree matrix.

In the embodiment of this application, the test feature matrix after dimensionality reduction

The historical feature matrix after each dimensionality reduction

The test platform calculates the test feature matrix after dimensionality reduction through the cosine theorem

The dimensionality-reduced feature of the n-th front-end test case in , and the historical feature matrix of the k-th dimensionality-reduction matrix

The similarity of the dimensionality-reduced features of each front-end historical case in

based on multiple similarities

Get the similarity matrix

Among them, the similarity matrix

The size is 1×N.

In a realizable application scenario, the test platform calculates the dimensionality-reduced test feature matrix through the cosine theorem

The nth row in and the historical feature matrix after dimensionality reduction

The similarity between each column in

Step 309 , based on the similarity matrix, when it is determined that all front-end test cases satisfy the automated test conditions, perform automated tests on all front-end test cases in the system version set to be tested.

In some embodiments, step 309 is based on the similarity matrix, and when it is determined that all front-end test cases meet the automated test conditions, all front-end test cases in the system version set to be tested are automatically tested, which can be implemented through the steps shown in Figure 7:

Step 601. Obtain the weight corresponding to each historical system version set in at least two historical system version sets.

In the embodiment of the present application, the test platform sets a weight w _k for the similarity matrix θ _k corresponding to each historical system version set, and the weight w _k is greater than or equal to 0 and less than or equal to 1. Here, when setting the weight w _k for each similarity matrix θ _k , it can be obtained by the following formula,

Among them, w _k is the weight w _k corresponding to each similarity matrix θ _k , k is the sequence number of each historical system version set in all historical system version sets, and K is the total number of all historical system version sets; what needs to be explained is , the smaller the sorting number k of each historical system version set, the similarity matrix θ _k corresponding to each historical system version set, the larger the weight w _k is set, the historical system corresponding to the minimum value k of the system under test The greater the possibility of continuous incremental development in an iterative manner.

In the embodiment of the present application, the test platform obtains the weight w _k corresponding to each historical system version set k in at least two historical system version sets.

Step 602, based on the similarity matrix between each front-end test case after dimension reduction and case n in all front-end historical cases in each historical version set after dimensionality reduction, and the weight corresponding to each historical system version set, generate Each target correlation matrix between the version set of the system under test and the set of all historical system versions.

In the embodiment of the present application, each front-end test case after dimension reduction includes the feature after dimension reduction of the front-end test case, and the cases in all front-end historical cases in each historical version set after dimension reduction include the features of each front-end historical case. Features after dimensionality reduction.

In the embodiment of this application, the test platform is based on the similarity matrix between each front-end test case after dimensionality reduction and case n in all front-end historical cases in each historical version set after dimensionality reduction

Compose the target similarity matrix

Based on the target similarity matrix θ _k and the weight w _k corresponding to each historical system version set k, the test platform generates each target between the system version set to be tested and all historical system version sets by y _k =w _k θ _k Incidence matrix y _k .

In other embodiments of the present application, the test platform generates each target correlation matrix y _k between the system version set to be tested and all historical system version sets, which can also be implemented in the following manner:

Step 1. Acquire weights and supplementary factors corresponding to each historical system version set in at least two historical system version sets.

Step2, based on the similarity matrix between each front-end test case after dimensionality reduction and case n in all front-end historical cases in each historical version set after dimensionality reduction, the weight and supplementary factor corresponding to each historical system version set , to generate each target correlation matrix between the version set of the system under test and the set of all historical system versions.

In the embodiment of this application, the test platform obtains the weight w _k corresponding to each historical system version set in at least two historical system version sets, and after supplementing the factor h, compares each front-end test case after dimensionality reduction with the reduced dimensionality The similarity matrix θ _k between cases n in all front-end historical cases in each historical version set, the weight w _k corresponding to each historical system version set and the supplementary factor h, through y _k =w _k θ _k +h, Generate each target correlation matrix y _k between the version set of the system to be tested and the set of all historical system versions. In this way, by setting the supplementary factor h, each element in each target correlation matrix y _k between the generated system-under-test version set and all historical system version sets is prevented from being 0.

Step 603: Obtain the maximum value of each row in each target correlation matrix, and determine a first number of maximum values greater than the first target threshold among the maximum values of all rows in each target correlation matrix.

In the embodiment of the present application, the test platform obtains the maximum value of each row in each target correlation matrix y _k , and obtains the maximum value greater than the first target threshold such as 1/2 among the maximum values of all rows in each target correlation matrix y _k , and determine the first quantity sum1 greater than the maximum value of 1/2.

Step 604. Calculate the ratio of the first quantity to the total number of all front-end test cases in the system version set under test to obtain the first ratio.

In the embodiment of the present application, the test platform calculates the ratio of the first quantity sum1 to the total number N of all front-end test cases in the system version set to be tested to obtain the first ratio z1, wherein,

Step 605. Obtain a second number of first ratios greater than the second target threshold among all first ratios corresponding to each target correlation matrix.

In the embodiment of the present application, the test platform obtains the first ratio z1 greater than the second target threshold such as 1/2 among all the first ratios z1 corresponding to each target correlation matrix y _k , and determines the first ratio z1 greater than 1/2 The second quantity sum2.

Step 606. Calculate the ratio of the second quantity to the total number of the historical system version sets to obtain the second ratio.

In the embodiment of the present application, the test platform calculates the ratio of the second number sum2 to the total number K of the historical system version set to obtain the second ratio z2, wherein,

Step 607. If the second ratio is greater than the third target threshold, determine that all front-end test cases meet the automated test conditions, and perform automated tests on all front-end test cases in the system version set to be tested.

In the embodiment of the present application, the test platform determines that the second ratio z2 is greater than the third target threshold such as 1/2, which means that all front-end test cases in the system version set to be tested are compared with all front-end historical test cases in the historical system version set. The difference is not big. At this point, it is determined that all front-end test cases meet the automated test conditions. And after the developer makes a small change to some of the scripts in the automation scripts corresponding to the historical system version set, the test platform can perform automated tests on all front-end test cases in the system version set to be tested through the modified automation scripts, so that It establishes a unified standard for the automated execution of front-end test cases, improves the accuracy of judgment, and improves processing efficiency without relying on manual operations.

Referring to FIG. 8, FIG. 8 is an optional flow chart of the test case processing method provided by the embodiment of the present application, which will be described in conjunction with the steps shown in FIG. 8,

Step 701. Obtain all test cases in the version set of the system to be tested.

Step 702: Input all test cases in the version set of the system to be tested into the trained classifier, and obtain all front-end test cases in all test cases output by the trained classifier.

Step 703. Obtain all front-end historical cases in each historical system version set in at least two historical system version sets.

Step 704 , based on the word frequency-reverse document frequency algorithm, calculate the eigenvalues corresponding to the features in each front-end test case in all front-end test cases, and obtain the test feature matrix of the system version set to be tested.

Step 705: Based on the word frequency-reverse document frequency algorithm, calculate the eigenvalues corresponding to the features in each front-end historical case in all front-end historical cases in each historical system version set, and obtain the historical feature matrix of each historical system version set.

Step 706 , based on the non-negative matrix factorization algorithm, perform dimension reduction processing on the test feature matrix to obtain a dimension-reduced test feature matrix.

Step 707 , based on the non-negative matrix factorization algorithm, perform dimensionality reduction processing on the historical feature matrix to obtain a dimensionally reduced historical feature matrix.

Step 708. Calculate the similarity between the dimensionality-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced feature of each front-end historical case in the dimension-reduced historical feature matrix, and obtain the similarity degree matrix.

Step 709, based on the obtained weight and supplementary factor corresponding to each historical system version set, and the relationship between each front-end test case after dimension reduction and case n in all front-end historical cases in each historical version set after dimension reduction The similarity matrix of is used to generate each target correlation matrix between the version set of the system under test and the set of all historical system versions.

Step 710, based on each target correlation matrix between the system version set to be tested and all historical system version sets, determine whether all front-end test cases meet the automated test conditions, so as to determine whether to perform all front-end test cases in the system version set to be tested automated test.

As can be seen from the above, in the embodiment of the present application, after the test platform obtains all front-end test cases and all front-end historical cases in the system version set to be tested, first, it calculates the number of front-end test cases in each front-end test case through the word frequency-reverse file frequency algorithm. The eigenvalues corresponding to the features of the system to be tested are obtained from the test feature matrix of the system version set to be tested, and the eigenvalues corresponding to the features in each front-end historical case are calculated to obtain the historical feature matrix of each historical system version set; secondly, through the non-negative matrix Algorithm, the dimensionality reduction process is performed on the test feature matrix and the historical feature matrix respectively, and the similarity processing is performed on the dimensionality-reduced test feature matrix and the dimensionality-reduced historical feature matrix to obtain a similarity matrix, and then according to the similarity matrix, When it is determined that all front-end test cases meet the automation test conditions, perform automated tests on all front-end test cases. In this way, this application solves the problem that related technologies must rely on manual operation and uncertainty brought by manual subjectivity, and at least has the problems of low efficiency and poor accuracy; realizes the establishment of a unified standard for the automatic execution of front-end test cases, and improves The accuracy of judgment is improved, and at the same time, it does not need to rely on manual operation, which improves the processing efficiency.

The following continues to illustrate the exemplary structure of the test case processing device 154 provided by the embodiment of the present application implemented as a software module. In some embodiments, as shown in FIG. 1 , the software modules stored in the test case processing device 154 of the memory 150 can be is a test case processing device in the test platform 100, including:

An acquisition module 1541, configured to acquire all front-end test cases in the system version set to be tested, and all front-end historical cases in each of the at least two historical system version sets;

The processing module 1542 is used to calculate the eigenvalues corresponding to the features in each front-end test case in all front-end test cases, and obtain the test feature matrix of the system version set to be tested;

The processing module 1542 is also used to calculate the eigenvalues corresponding to the features in each front-end historical case in all front-end historical cases in each historical system version set, and obtain the historical feature matrix of each historical system version set;

The processing module 1542 is also used to perform dimensionality reduction processing on the test feature matrix based on the non-negative matrix factorization algorithm, and obtain the test feature matrix after dimensionality reduction;

The processing module 1542 is also used to perform dimensionality reduction processing on the historical feature matrix based on the non-negative matrix factorization algorithm to obtain the historical feature matrix after dimensionality reduction;

The processing module 1542 is also used to calculate the similarity between the dimension-reduced features of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced features of each front-end historical case in the dimension-reduced historical feature matrix degree, to obtain a similarity matrix; wherein, n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the system version set to be tested;

The processing module 1542 is further configured to perform automated testing on all front-end test cases in the system version set to be tested when it is determined that all front-end test cases meet the automated test conditions based on the similarity matrix.

In some embodiments, the processing module 1542 is further configured to decompose the test feature matrix through a non-negative matrix decomposition algorithm based on the determined feature number after dimensionality reduction of the test feature matrix to obtain a test projection matrix and a test fundamental matrix; obtain Module 1541 is also used to obtain the first product matrix obtained by multiplying the test projection matrix and the test fundamental matrix; the processing module 1542 is also used to obtain the first difference matrix obtained by subtracting the first product matrix from the test feature matrix according to the difference value The threshold matrix is used to determine the test projection matrix corresponding to the first difference matrix as the dimension-reduced test feature matrix.

In some embodiments, the processing module 1542 is further configured to adjust each element in the test projection matrix through the projection matrix adjustment model to obtain the adjusted test projection matrix if the first difference matrix does not meet the difference threshold matrix ; Each element in the test fundamental matrix is adjusted through the fundamental matrix adjustment model to obtain the adjusted test fundamental matrix; the acquisition module 1541 is also used to obtain the adjusted test projection matrix and multiply the adjusted test fundamental matrix to obtain The second product matrix; the processing module 1542 is also used to determine the adjusted test projection matrix corresponding to the second difference matrix if the second difference matrix obtained by subtracting the second product matrix from the test feature matrix meets the difference threshold matrix is the test feature matrix after dimensionality reduction.

In some embodiments, the projection matrix adjustment model is:

Among them, W' _ns is the element in row n and column s in the adjusted test projection matrix, W is the test projection matrix, W _ns is the element in row n and column s in the test projection matrix, X is the test feature matrix, B is the test fundamental matrix, B ^T is the transpose matrix of the test fundamental matrix, (XB ^T ) _ns is the nth row s of the matrix obtained by multiplying the test feature matrix X and the transpose matrix B ^T of the test fundamental matrix The element of the column, (WBB ^T ) _ns is the element of the nth row and the sth column in the matrix obtained after multiplying the transposition matrix B ^T of the test projection matrix W, the test fundamental matrix B and the test fundamental matrix;

The basic matrix adjustment model is:

In some embodiments, the processing module 1542 is further configured to adjust each element in the adjusted test projection matrix through the projection matrix adjustment model to obtain a new adjusted The test projection matrix; each element in the adjusted test fundamental matrix is adjusted through the fundamental matrix adjustment model to obtain a new adjusted test fundamental matrix; the acquisition module 1541 is also used to obtain the newly adjusted test projection matrix and The third product matrix obtained by multiplying the newly adjusted test fundamental matrix; the processing module 1542 is also used to determine the third difference matrix if the third difference matrix obtained by subtracting the third product matrix from the test feature matrix meets the difference threshold matrix. The newly adjusted test projection matrix corresponding to the value matrix is the dimensionality-reduced test feature matrix.

In some embodiments, the obtaining module 1541 is also used to obtain the weight corresponding to each historical system version set in at least two historical system version sets; the processing module 1542 is also used to obtain the weight corresponding to each front-end test case and The similarity matrix between cases n in all front-end historical cases in each historical version set after dimensionality reduction, and the weight corresponding to each historical system version set, generate the relationship between the system version set to be tested and all historical system version sets Each target correlation matrix; the acquisition module 1541 is also used to obtain the maximum value of each row in each target correlation matrix, and determine among the maximum values of all rows in each target correlation matrix, which is greater than the maximum value of the first target threshold The first quantity; the processing module 1542 is also used to calculate the ratio of the first quantity and the total number of all front-end test cases in the system version set to be tested to obtain the first ratio; the acquisition module 1541 is also used to obtain each target correlation matrix Among all the corresponding first ratios, the second number of the first ratio greater than the second target threshold; the processing module 1542 is also used to calculate the ratio of the second number to the total number of historical system version sets to obtain the second ratio; if the second If the second ratio is greater than the third target threshold, it is determined that all front-end test cases meet the automated test conditions, and automated tests are performed on all front-end test cases in the system version set to be tested.

In some embodiments, the obtaining module 1541 is also used to obtain all test cases in the version set of the system under test; the processing module 1542 is also used to input all test cases in the version set of the system under test to the trained classifier In , get all front-end test cases in all test cases output by the trained classifier.

The embodiment of the present application provides a storage medium storing executable instructions, and the executable instruction is stored therein. When the executable instruction is executed by the processor, it will cause the processor to execute the method provided in the embodiment of the present application, for example, as shown in FIG. 2 - the method shown in Figure 3, Figure 5-Figure 8.

In some embodiments, the storage medium can be a computer-readable storage medium, for example, a ferroelectric memory (FRAM, Ferromagnetic Random Access Memory), a read-only memory (ROM, Read Only Memory), a programmable read-only memory (PROM, Programmable Read Only Memory), Erasable Programmable Read Only Memory (EPROM, Erasable Programmable Read Only Memory), Electrically Erasable Programmable Read Only Memory (EEPROM, Electrically Erasable Programmable Read Only Memory), flash memory, magnetic surface memory, optical disc, Or memory such as CD-ROM (Compact Disk-Read Only Memory); It can also be various devices including one or any combination of the above-mentioned memories.

In some embodiments, executable instructions may take the form of programs, software, software modules, scripts, or code written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and its Can be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

As an example, executable instructions may, but do not necessarily correspond to files in a file system, may be stored as part of a file that holds other programs or data, for example, in Hyper Text Markup Language (HTML, Hyper Text Markup Language) in one or more scripts in a document, in a single file dedicated to the program in question, or in multiple cooperating files (for example, a file that stores one or more modules, subroutines, or code sections )middle. As an example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices located at one site, or alternatively, on multiple computing devices distributed across multiple sites and interconnected by a communication network. to execute.

The above descriptions are merely examples of the present application, and are not intended to limit the protection scope of the present application. Any modifications, equivalent replacements and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Industrial Applicability

The embodiment of the present application provides a test case processing method, test case processing device, test platform and storage medium, by obtaining all front-end test cases in the system version set to be tested, and each historical system in at least two historical system version sets All front-end historical cases in the version set; calculate the eigenvalues corresponding to the features in each front-end test case in all front-end test cases, and obtain the test feature matrix of the system version set to be tested; calculate all front-end history in each historical system version set In the case, the eigenvalues corresponding to the features in each front-end historical case are obtained to obtain the historical feature matrix of each historical system version set; based on the non-negative matrix decomposition algorithm, the dimensionality reduction process is performed on the test feature matrix to obtain the dimensionality-reduced test features matrix; based on the non-negative matrix factorization algorithm, the dimensionality reduction process is performed on the historical feature matrix to obtain the dimensionality-reduced historical feature matrix; the dimensionality-reduced feature and the dimensionality-reduced Based on the similarity of the dimensionality-reduced features of each front-end historical case in the dimensioned historical feature matrix, a similarity matrix is obtained; based on the similarity matrix, when it is determined that all front-end test cases meet the automated test conditions, the test system version set All front-end test cases are automatically tested; that is to say, after the test platform obtains all front-end test cases and all front-end historical cases in the system version set to be tested, first, it calculates the feature value corresponding to the feature in each front-end test case , to obtain the test feature matrix of the system version set to be tested, calculate the eigenvalues corresponding to the features in each front-end historical case, and obtain the historical feature matrix of each historical system version set; secondly, through the non-negative matrix algorithm, the test feature matrix and the historical feature matrix are respectively subjected to dimensionality reduction processing, and the similarity processing is performed on the dimensionality-reduced test feature matrix and the dimensionality-reduced historical feature matrix to obtain a similarity matrix, and then according to the similarity matrix, it is determined that all front-end test cases satisfy When automating test conditions, automate testing for all front-end test cases. In this way, this application solves the problem that related technologies must rely on manual operation and uncertainty brought by manual subjectivity, and at least has the problems of low efficiency and poor accuracy; realizes the establishment of a unified standard for the automatic execution of front-end test cases, and improves The accuracy of judgment is improved, and at the same time, it does not need to rely on manual operation, which improves the processing efficiency.

Claims

A test case processing method, comprising:

Obtain all front-end test cases in the system version set to be tested, and all front-end historical cases in each historical system version set in at least two historical system version sets;

Calculate the eigenvalues corresponding to the features in each front-end test case in all the front-end test cases, and obtain the test feature matrix of the version set of the system to be tested;

Calculating the eigenvalues corresponding to the features in each of the front-end historical cases in all the front-end historical cases in the set of each historical system version, to obtain the historical feature matrix of each set of historical system versions;

Based on a non-negative matrix factorization algorithm, performing dimensionality reduction processing on the test feature matrix to obtain a dimensionally reduced test feature matrix;

Based on the non-negative matrix decomposition algorithm, performing dimensionality reduction processing on the historical feature matrix to obtain a dimensionally reduced historical feature matrix;

Calculating the similarity between the dimension-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and the dimension-reduced feature of each front-end historical case in the dimension-reduced historical feature matrix, to obtain A similarity matrix; wherein, the n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the version set of the system to be tested;

Based on the similarity matrix, when it is determined that all the front-end test cases meet the automated test conditions, the automated test is performed on all the front-end test cases in the version set of the system under test.
The method according to claim 1, wherein the non-negative matrix factorization algorithm is used to perform dimensionality reduction processing on the test feature matrix to obtain a dimensionally reduced test feature matrix, including:

Based on the determined eigennumbers after dimension reduction of the test feature matrix, the test feature matrix is decomposed by the non-negative matrix factorization algorithm to obtain a test projection matrix and a test fundamental matrix;

Obtaining a first product matrix obtained by multiplying the test projection matrix and the test fundamental matrix;

If the first difference matrix obtained by subtracting the first product matrix from the test feature matrix conforms to the difference threshold matrix, determine that the test projection matrix corresponding to the first difference matrix is the dimension-reduced test feature matrix.
The method according to claim 2, wherein the method further comprises:

If the first difference matrix does not conform to the difference threshold matrix, each element in the test projection matrix is adjusted through a projection matrix adjustment model to obtain an adjusted test projection matrix;

Adjust each element in the basic test matrix through the basic matrix adjustment model to obtain an adjusted basic test matrix;

Obtaining a second product matrix obtained by multiplying the adjusted test projection matrix and the adjusted test fundamental matrix;

If the second difference matrix obtained by subtracting the second product matrix from the test feature matrix conforms to the difference threshold matrix, determine that the adjusted test projection matrix corresponding to the second difference matrix is the The reduced test feature matrix.
The method according to claim 3, wherein the projection matrix adjustment model is:

Wherein, the W' ns is the element of the nth row and the sth column in the adjusted test projection matrix, the W is the test projection matrix, and the W ns is the nth row in the test projection matrix The elements in the sth column, the X is the test feature matrix, the B is the test fundamental matrix, the B T is the transpose matrix of the test fundamental matrix, and the (XB T ) ns is the The elements of the nth row and the sth column in the matrix obtained after the test feature matrix X is multiplied by the transpose matrix B T of the test fundamental matrix, the (WBB T ) ns is the test projection matrix W, the The elements of the nth row and the sth column in the matrix obtained after the multiplication of the test fundamental matrix B and the transpose matrix B T of the test fundamental matrix;

The basic matrix adjustment model is:

Wherein, the B' sm is the element of the sth row and the mth column in the adjusted test fundamental matrix, the B is the test fundamental matrix, and the B sm is the sth row in the test fundamental matrix The element in the m column, the X is the test feature matrix, the W is the test projection matrix, the W T is the transpose matrix of the test projection matrix, and the (W T X) sm is The transpose matrix W T of the test projection matrix is multiplied by the test feature matrix X, and the elements in the sth row and the mth column of the matrix are obtained, and the (W T WB) sm is the element of the test projection matrix Elements in row s and column m in a matrix obtained by multiplying the transpose matrix W T , the test projection matrix W and the test fundamental matrix B.
The method according to claim 3, wherein the method further comprises:

If the second difference matrix does not conform to the difference threshold matrix, adjust each element in the adjusted test projection matrix through the projection matrix adjustment model to obtain a new adjusted test projection matrix;

Each element in the adjusted test fundamental matrix is adjusted by the fundamental matrix adjustment model to obtain a new adjusted test fundamental matrix;

Obtaining a third product matrix obtained by multiplying the newly adjusted test projection matrix and the newly adjusted test fundamental matrix;

If the third difference matrix obtained by subtracting the third product matrix from the test feature matrix conforms to the difference threshold matrix, determine that the newly adjusted test projection matrix corresponding to the third difference matrix is the reduced Dimensional test feature matrix.
The method according to claim 1, wherein, based on the similarity matrix, when it is determined that all the front-end test cases meet the automated test conditions, the automated test is performed on all the front-end test cases in the version set of the system under test ,include:

Obtain the weight corresponding to each historical system version set in at least two historical system version sets;

Based on the similarity matrix between each front-end test case after dimensionality reduction and case n in all front-end history cases in each historical version set after dimensionality reduction, each historical system version set Corresponding weights, generating each target correlation matrix between the system version set to be tested and all historical system version sets;

Obtaining the maximum value of each row in each target incidence matrix, and determining a first number of maximum values greater than a first target threshold among the maximum values of all rows in each target incidence matrix;

calculating the ratio of the first quantity to the total number of all front-end test cases in the system version set under test to obtain the first ratio;

Obtaining a second number of the first ratios greater than a second target threshold among all the first ratios corresponding to each target incidence matrix;

calculating the ratio of the second quantity to the total number of the historical system version set to obtain a second ratio;

If the second ratio is greater than the third target threshold, it is determined that all the front-end test cases meet the automated test conditions, and the automated test is performed on all the front-end test cases in the version set of the system under test.
The method according to any one of claims 1 to 6, wherein, before obtaining all front-end test cases in the system version set to be tested, the method further includes:

Obtain all test cases in the system version set to be tested;

Inputting all the test cases in the version set of the system to be tested into the trained classifier to obtain all the front-end test cases among the all test cases output by the trained classifier.
A test case processing device, wherein the device comprises:

Obtaining module, used to obtain all front-end test cases in the system version collection to be tested, and all front-end historical cases in each historical system version collection in at least two historical system version collections;

A processing module, configured to calculate the eigenvalues corresponding to the features in each of the front-end test cases in all the front-end test cases, and obtain the test feature matrix of the version set of the system to be tested;

The processing module is also used to calculate the feature value corresponding to the feature in each front-end historical case in all the front-end historical cases in the set of each historical system version, and obtain the historical feature of each historical system version set matrix;

The processing module is also used to perform dimensionality reduction processing on the test feature matrix based on a non-negative matrix factorization algorithm to obtain a dimensionality-reduced test feature matrix;

The processing module is further configured to perform dimensionality reduction processing on the historical feature matrix based on the non-negative matrix factorization algorithm, to obtain a dimensionally reduced historical feature matrix;

The processing module is also used to calculate the dimensionality-reduced feature of the nth front-end test case in the dimension-reduced test feature matrix and the dimensionality reduction of each front-end historical case in the dimension-reduced historical feature matrix The degree of similarity of the features after is obtained similarity matrix; Wherein, said n is a positive integer greater than or equal to 1 and less than or equal to N, and N is the total number of front-end test cases in the described system version set to be tested;

The processing module is further configured to perform automated testing on all front-end test cases in the version set of the system under test when it is determined that all front-end test cases meet the automated test conditions based on the similarity matrix.
A test platform, including:

The memory is used to store executable instructions; the processor is used to implement the method according to any one of claims 1 to 7 when executing the executable instructions stored in the memory.
A storage medium, wherein executable instructions are stored for causing a processor to implement the method according to any one of claims 1 to 7.