WO2017198198A1

WO2017198198A1 - Software compiling method and device

Info

Publication number: WO2017198198A1
Application number: PCT/CN2017/084942
Authority: WO
Inventors: 徐磊
Original assignee: 中兴通讯股份有限公司
Priority date: 2016-05-19
Filing date: 2017-05-18
Publication date: 2017-11-23
Also published as: CN107402797A

Abstract

A software version compiling method and apparatus. The method comprises: according to a pre-established failed creation information repository, determining a target frequent item set in the repository (S10), wherein an item set for each previous failed creation is stored in the failed creation information repository; calculating a dependence relationship between modules according to the target frequent item set, and determining an effective strong correlation rule (S20); and binding together the modules which are in an effective strong correlation, to simultaneously compile orsimultaneously give an early warning (S30).

Description

Software compilation method and device

Technical field

The present disclosure relates to the field of software technologies, for example, to a software compilation method and apparatus.

Background technique

Software version compilation is an indispensable link in continuous integration. For large software systems, such as communication systems, there are often many modules, and the degree of association between modules is high. It is easy to take the whole system, often because of a module. Small modifications caused the entire system to fail to compile. Because there are many modules involved, the involved personnel are extensive, the compilation troubleshooting cycle is long, and the severe blocking continues to integrate, resulting in a high risk of software version update.

When the version compilation fails, it is often relied on the maintenance personnel to solve the problem one by one to solve the problem of failure to compile. On the one hand, the maintenance personnel will be trapped in the maintenance and fault location of the basic environment for a long time, and compile at the same time. In the case of fault location, the influence of inter-dependency between multiple modules is easily ignored, the troubleshooting is blocked, and the fault resolution period is long. On the other hand, it is difficult to find a professional who understands the dependencies of all modules and modules to perform troubleshooting.

Summary of the invention

The present disclosure provides a software compiling method and apparatus, which can solve the problem that the manual solving of the compiling problem is difficult, and the problem of inter-module interdependence is not considered when compiling the fault.

A method for compiling a version software provided by this embodiment, the method may include: determining, according to a pre-established failure information construction knowledge base, a target frequent item set in the knowledge base, wherein the failure construction information knowledge base The item set that was previously failed to be saved is saved; the effective strong association rule is determined according to the dependency relationship between the target frequent item set calculation modules; and the modules of the effective strong association are bound together to compile or simultaneously warn.

Optionally, the method may further include: acquiring, when each compilation fails, acquiring the item set of the failed build and Add to the knowledge base.

Optionally, determining the target frequent item set in the knowledge base according to the pre-established failure building information knowledge base may include: sequentially calculating a frequent item set of each order according to the item set in the knowledge base Calculating the support degree of each 1st order item set in the knowledge base, pruning the 1st order item set whose support degree is lower than the first threshold, and obtaining the 1st order frequent item set; according to the m-1 order frequent item set, Using lexicographic order and combination method, multiple m-1 order frequent cameras are generated to generate multiple m-th order items, and the support degree of each m-th order item set is calculated respectively, and the support degree is lower than the preset corresponding to the preset m-th order item set. The m-th order item set with the support threshold is set for pruning, and the m-th order frequent item set is obtained, and when the number of m-th order frequent cameras is one, the m-th order frequent item set is determined as the target frequent item set, wherein m is A positive integer and m is greater than 1.

Optionally, the following formulas are used to calculate the support of multiple m-th order frequent items:

Support(X)=P(X)/P(I);

Support(X->Y)=P(X∪Y)/P(I);

Where I represents the total item set in the knowledge base, Support(X) represents the probability of the item set {X} appearing in the total item set, and Support(X->Y) indicates that the item set {X, Y} is in the total item. The probability of occurrence in the set, P represents the probability, P(X)/P(I) represents the probability of the item set (X) appearing in the total item set, and the support degree of the item set means that the item set appears in the total item set The probability.

Optionally, determining the valid strong association rule according to the dependency relationship between the target frequent item set calculation modules may include: calculating a confidence level of each non-empty item set in the target frequent item set, which is lower than The non-empty item set corresponding to the confidence level of the pre-set confidence threshold is pruned to obtain a plurality of target non-empty item sets; and the plurality of target non-empty item sets are calculated according to the confidence of the plurality of target non-empty item sets The degree of lift; and the rule between multiple target non-empty item sets that determine a degree of lift greater than one is a valid strong association rule.

Optionally, the calculating the confidence of each non-empty item set in the target frequent item set adopts the following formula:

Confidence(X->Y)=P(Y|X)=P(X∪Y)/P(X);

Among them, the confidence Confidence (X->Y) indicates that in the case of having the item set X, the probability of having the item set Y is derived by the association rule "X->Y".

Optionally, calculating the plurality of target non-empty items according to the confidence of the plurality of target non-empty item sets For the degree of improvement between each other, the following formula is used:

Lift(X->Y)=P(Y|X)/P(Y)=P(X∪Y)/P(Y);

Wherein, the lift Lift(X->Y) represents the ratio of the probability of containing the item set Y under the condition that the item set X is included, and the probability of containing the item set Y under the condition that the item set X is not included;

If Lift(X->Y)>1, it means that the rule X->Y is a valid strong association rule;

If Lift(X->Y)<=1, it means that the rule X->Y is an invalid strong association rule;

If Lift(X->Y)=1, it means that item set X and item set Y are independent of each other.

The embodiment further provides a version software compiling device, which may include: a target frequent item set determining unit, configured to determine a target frequent item set in the knowledge base according to a pre-established failure building information knowledge base, wherein The failed build information knowledge base stores the item set that is previously failed to be built; the effective strong association rule analysis unit is configured to determine the effective strong association rule according to the dependency relationship between the target frequent item set calculation modules; and the associated compilation The Build Unit is set to bind the modules that are effectively strongly associated together and compile at the same time.

Optionally, the target frequent item set determining unit may include: a first determining subunit, configured to sequentially calculate a frequent item set of each order according to the item set of the knowledge base; and second determining the subunit, configured as a calculation center Determining the support degree of each 1st order item set in the knowledge base, pruning the 1st order item set whose support degree is lower than the first threshold, and obtaining the 1st order frequent item set; and the third determining subunit, which is set according to m- The first-order frequent itemsets are lexicographically ordered and combined, and multiple m-1 order frequent itemsets are generated to generate multiple m-th order items, and the support degrees of multiple m-th order items are calculated respectively. The support degree is lower than m order. The m-th order item set of the preset support degree threshold corresponding to the item set is pruned, and the m-th order frequent item set is obtained. When the number of the m-th order frequent item set is one, the m-th order frequent item set is determined as the target frequent item set. Stop calculating the m+1th order item set, where m is a positive integer and m is greater than 1.

Optionally, the effective strong association rule analysis unit may include: a first analysis subunit, configured to calculate a confidence level of each non-empty item set in the target frequent item set, and a non-empty item lower than a preset reliability threshold The set performs pruning to obtain a plurality of target non-empty item sets; and the second analysis sub-unit is configured to calculate a degree of elevation between the plurality of target non-empty item sets according to the confidence of the plurality of target non-empty item sets; The third analysis subunit is set to determine that the rule between the plurality of target non-empty item sets having a degree of lift greater than 1 is effective Association rules.

The embodiment further provides a computer readable storage medium storing computer executable instructions for executing any of the above version software compilation methods.

The embodiment also provides a server including one or more processors, a memory, and one or more programs, the one or more programs being stored in the memory, when executed by one or more processors, Execute any of the above versions of the software compilation method.

The embodiment further provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions, when the program instructions are executed by a computer And causing the computer to execute any of the above versions of the software compilation method.

The version software compiling method and apparatus provided by the present disclosure automatically calculates the degree of association between modules according to a pre-established failure to construct an information knowledge base, and determines effective strong association rules; and binds the modules of effective strong association according to the effective strong association rules. Compile at the same time or at the same time. It is the most likely to achieve self-repair of the compiled project, ensuring continuous integration and reducing the risk of release. It can also save the wrong human resources when the compilation fails.

DRAWINGS

FIG. 1 is a flowchart of a method for compiling a version software according to an embodiment of the present invention.

FIG. 2 is a flowchart of a method for determining a target frequent item set according to an embodiment of the present disclosure.

FIG. 3 is a flowchart of an effective strong association rule analysis method according to an embodiment of the present invention.

FIG. 4 is a schematic structural diagram of a version software compiling apparatus according to an embodiment of the present invention.

FIG. 5 is a schematic structural diagram of a target frequent item set determining unit according to an embodiment of the present disclosure.

FIG. 6 is a schematic structural diagram of an effective strong association rule analysis unit according to an embodiment of the present disclosure.

FIG. 7 is a schematic structural diagram of a general hardware of a server according to an embodiment of the present disclosure.

detailed description

The present disclosure will be described below in conjunction with the accompanying drawings and embodiments. It should be understood that the description described here The examples are merely illustrative of the disclosure and are not intended to limit the disclosure. The technical features in the following embodiments and embodiments may be combined with each other without conflict.

The following describes an embodiment of the version software compiling method in this embodiment. The execution body of the version software compiling method is a version software compiling device, and the software compiling device may be located in a server or a terminal.

Referring to FIG. 1, the version software compiling method in this embodiment may include S10-S30.

In S10, a target frequent item set in the knowledge base is determined according to a pre-established failure building information knowledge base.

Optionally, the failed build information knowledge base saves the item set that was previously failed to be built, and each compiled failure module corresponds to one item set. For example, I={i1, i2,..., ij} is a collection of j different items, and one element in the set is called an item. The set I of items is called an item set, the number of elements is called the length of the item set, and the item set of length k is called the k-order item set. For example, the item set is I={A, B, C, D, E, F}, and the length of I is 6. The set of tasks that are failed to build each time is a subset of the total item set I.

Optionally, referring to FIG. 2, in S10, determining a target frequent item set in the knowledge base includes S101-S103 according to the pre-established failure building information knowledge base.

In S101, a frequent item set of each order is sequentially calculated according to the item set of the knowledge base.

Optionally, a preset support degree threshold is preset for each order item, and the preset support degree threshold may represent a minimum importance of the association rule. A frequent item set may refer to a set of items whose support degree is not less than a preset support degree threshold corresponding to the order item set. A frequent item set of length k can be referred to as a k-th order frequent item set.

Optionally, the item set of each order corresponds to a preset support degree threshold of the order, and the item set pruning whose support degree is lower than the preset support threshold of the order is supported, and the support degree is greater than or equal to the preset support of the order. The set of items of the degree threshold constitutes a frequent item set of the order.

Optionally, calculating the first-order frequent item set includes: calculating a support degree of each first-order item set in the knowledge base, and performing pruning on the first-order item set whose support degree is lower than the first threshold, and supporting the degree A set of 1st order items greater than or equal to a preset support degree threshold of the 1st order item set constitutes a 1st order frequent item set.

Optionally, a set of items that are less than a preset support threshold of the set of 1st order items are ignored. Pruning means that the set of items that are less than the preset support threshold of the first-order item set is ignored.

Optionally, in this embodiment, the following formula may be used to calculate the support degree of the n-th order frequent itemsets:

Support(X)=P(X)/P(I);

Support(X->Y)=P(X∪Y)/P(I);

Where I represents the total item set in the knowledge base, Support(X) represents the probability of the item set {X} appearing in the total item set, and Support(X->Y) indicates that the item set {X, Y} is in the total item. The probability of occurrence in the set.

For example, in the compilation task of four builds, the following table 1 module item set (referred to as item set) fails.

Table 1

构建次数Number of builds	构建失败的模块项集Build a failed module item set
11	A，C，DA, C, D
22	B，C，EB, C, E
33	A，B，C，EA, B, C, E
44	B，EB, E

In S102, a single module item set, that is, a 1st order item set is calculated, and the first degree item set lower than the preset support degree threshold corresponding to the 1st order item set is pruned in the total item set support degree.

Optionally, it is assumed that the preset support degree threshold (which may be referred to as a first threshold) corresponding to the first-order item set is 50%. As shown in Table 2 below, the knowledge base item set includes four first-order item sets, and each first-order item is calculated. The corresponding support degree is set, and the first-order item set whose support degree is lower than the preset support degree threshold is pruned (as shown in FIG. 2, the first-order item set {D} is pruned), and the support degree is greater than or A first-order item set equal to a preset support degree threshold corresponding to the first-order item set constitutes a first-order frequent item set {A}, {B}, {C}, and {E}.

Table 2

In S103, according to the m-1 order frequent item set, the m-th order item set is generated by using the dictionary order and the combination manner, and the support degree of each m-th order item set is calculated respectively, and the preset with the support degree lower than the m-th order item set is preset. The m-th order item set of the support threshold is pruned to obtain an m-th order frequent item set until the last remaining high-order frequent item set.

Where m is a positive integer and m is greater than 1.

The so-called dictionary order, also known as Lexicographical Order, is a sorting method for forming sequences of random variables. Multiple random variables are compared one by one from left to right, and are arranged in order from small to large to form an ordered queue.

For example, the first-order frequent itemsets {A}, {B}, {C}, and {E} are lexicographically ordered, and the process of generating a second-order item set may be: combining four first-order frequent itemsets into two. The set of order terms is {A, B}, {B, C}, {C, E}, {A, C}, {B, E} and {A, E}. At this time, a plurality of second-order item sets are randomly arranged, and the plurality of second-order item sets are sorted in a lexicographic order, and first, the first item sets in the plurality of second-order item sets are respectively compared by two, and The plurality of second-order item sets are reordered in order from small to large, and the second item set in the plurality of second-order item sets after reordering is compared two by two, and again in order from small to large Multiple 2nd order items are sorted. Thus, the lexicographically ordered 2nd order item set is: {A, B}, {A, C}, {A, E}, {B, C}, {B, E}, {C, E}.

Based on the above examples, the obtained first-order frequent itemsets are {A}, {B}, {C}, and {E}. The first-order frequent itemsets are generated in a lexicographic order and a combination of two and two to generate a second-order item set {A, B}, {A, C}, {A, E}, {B, C}, {B, E }, {C, E}. The second-order item set whose support degree is lower than the preset support degree threshold (preset threshold value, for example, 25%) corresponding to the second-order item set is pruned, and the support degree is greater than or equal to the second-order item of the small support degree threshold. The set constitutes a second-order frequent item set {A, C}{B, C}, {B, E}{C, E}, as shown in Table 3 below.

table 3

The second-order frequent itemsets {A, C}, {B, C}, {B, E} and {C, E} obtained above are generated in a combined manner and in a lexicographic order to generate a third-order item set {A , B, C}, {A, C, E}, {A, B, E} and {B, C, E}. The third-order item set (for example, 25%) whose support degree is lower than the preset support degree threshold corresponding to the third-order item set is pruned, and the support degree is greater than or equal to the preset support degree threshold corresponding to the third-order item set. The set of order items constitutes the 3rd order frequent itemsets {B, C, E}, as shown in Table 4 below.

Table 4

Since there is only one third-order frequent item set {B, C, E}, the third-order frequent item set obtained so far is the target frequent item set.

In this embodiment, if there is only one first-order frequent item set, the first-order frequent item set is the target frequent item set, and if the first-order frequent item set is more than one, according to the obtained first order Frequent itemsets, using a lexicographical order and a two-two combination method, generating a 2nd-order item set, respectively calculating the support degree of each 2nd-order item set, and pruning the 2nd-order item set whose support degree is lower than the second threshold, A 2nd order item set whose support degree is greater than or equal to the second threshold value constitutes a 2nd order frequent item set.

If there is more than one second-order frequent item set, the third-order item set is continuously calculated, and the third-order frequent item set is calculated until the last remaining high-order frequent item set.

If the second-order frequent item set is more than one, continue to calculate the third-order item set, and the calculated multiple 3 The support degree of the order item set is compared with the preset support degree threshold corresponding to the 3rd order item set. If the support degree of all the 3rd order item sets is less than the preset support degree threshold corresponding to the 3rd order item set, it is determined that the requirement is not met. The frequent item set, the version software compilation process in this embodiment ends.

In S20, the effective strong association rule is determined according to the dependency relationship between the target frequent item set calculation modules.

Referring to FIG. 3, in S20, determining a valid strong association rule may include S201-S203 according to a dependency relationship between the target frequent item set calculation modules.

In S201, a confidence level of each non-empty item set in the target frequent item set is calculated, and a non-empty item set that does not satisfy the pre-set reliability threshold is pruned.

Optionally, when the confidence level of the non-empty item set is lower than the preset reliability threshold, determining that the confidence of the non-empty item set does not satisfy the pre-set confidence threshold; when the confidence of the non-empty item set is greater than or equal to When the reliability threshold is preset, it is determined that the execution degree of the non-empty item set satisfies the preset reliability threshold.

In S202, the degree of lifting between the non-empty item sets is calculated according to the confidence of the non-empty item set left after the pruning.

In S203, it is determined that the rule between the non-empty item sets whose degree of lift is greater than 1 is a valid strong association rule.

Optionally, the calculating the confidence of each non-empty item set in the target frequent item set may adopt the following formula:

Confidence(X->Y)=P(Y|X)=P(X∪Y)/P(X);

Optionally, the calculating the degree of lifting between the non-empty item sets according to the confidence of the non-empty item set left after the pruning may be performed by using the following formula:

Lift(X->Y)=P(Y|X)/P(Y)=P(X∪Y)/P(Y);

Among them, the lift degree Lift(X->Y) represents the ratio of the probability of containing the item set Y under the condition that the item set X is included, and the probability of containing the item set Y under the condition that the item set X is not included.

A rule that satisfies a preset support threshold and a preset reliability threshold is called a strong association rule. In a strong association rule, a valid strong association rule and an invalid strong association rule can be divided.

When Lift(X->Y)>1, it means that the rule X->Y is a valid strong association rule; when Lift(X->Y)<=1, it means that the rule X->Y is invalid. Association rule; when Lift(X->Y)=1, it means that item set X and item set Y are independent of each other.

Continuing with the above example, the target frequent item set is {B, C, E}, and the confidence between the individual non-empty item sets in the target frequent item set is calculated respectively, assuming that the pre-set confidence threshold is 75%, when a single non-empty item Rules with confidence that the set is less than the pre-set confidence threshold will be pruned, as shown in Table 5 below:

table 5

At this point, leave the non-empty item set B, E.

Continue to calculate the degree of lifting between non-empty item sets based on the confidence of the non-empty item set left after pruning, as calculated below.

Lift(B->E)=100%/75%=1.33;

Lift(E->B)=100%/75%=1.33;

At this point, valid strong association rules B->E and E->B are obtained.

In S30, the modules that are effectively strongly associated are bound together to compile or simultaneously warn.

Optionally, the modules that are effectively strongly associated are combined and compiled at the same time to maximize self-healing, and the modules with effective and strong associations are combined and early warning can reduce the human resources of manual troubleshooting and improve efficiency.

In step S20, valid strong association rules B->E and E->B are obtained. When the E construction fails, B can be constructed at the same time, or module B can be notified at the same time; when the E construction fails, it can be simultaneously B linkage construction, can also notify module E at the same time, so that B and E are associated, considering the relationship between different modules, version compilation according to the effective strong association rules, can be the most self-healing, to ensure continuous integration The effect is to reduce the risk of version release. Can warn when it is not self-healing Improve the efficiency of manual troubleshooting.

It should be noted that, in this embodiment, each time the compilation fails, the item set of the failed build is acquired and added to the knowledge base. Therefore, the knowledge base can be continuously updated, and the updated knowledge base can re-correct the effective strong association rules.

The version software compiling method provided in this embodiment automatically calculates the association degree between the modules by constructing the information knowledge base according to the pre-established failure, and determines the effective strong association rule; and binds the effective strong association modules together according to the effective strong association rule. Compile or alert at the same time. It can not only self-repair the compiled project, but also ensure continuous integration and reduce the risk of version release. It can also save the wrong human resources when the compilation fails.

The following describes an embodiment of the version software compiling apparatus in the embodiment of the present invention.

Referring to FIG. 5, it is a schematic diagram of an embodiment of a version software compiling apparatus in this embodiment. The version apparatus may include a target frequent item set determining unit 10, an effective strong association rule analyzing unit 20, and an associated compiling building unit 30.

The target frequent item set determining unit 10 is configured to determine a target frequent item set in the knowledge base according to the pre-established failure building information knowledge base, wherein the failed build information knowledge base stores the item set that was previously failed to be built. .

Optionally, the target frequent item set determining unit 10 may include: a first determining subunit 101, configured to sequentially calculate a frequent item set of each order according to the item set of the knowledge base; and second determining the subunit 102, setting To calculate the support degree of each 1st order item set in the knowledge base, pruning the 1st order item set whose support degree is lower than the first threshold to obtain a 1st order frequent item set; and the third determining subunit 103, setting According to the m-1 order frequent itemsets, the m-order itemsets are generated in lexicographic order, and the support degree of each m-th order item set is calculated separately, and the m-th order item set whose support degree is lower than the preset threshold is pruned to obtain m. The order frequent itemsets until the last remaining high-order frequent itemsets, where m is a positive integer and m is greater than 1.

Optionally, the third determining subunit 103 may be configured to calculate an nth order frequent item set support degree by using the following formula:

Support(X)=P(X)/P(I);

Support(X->Y)=P(X∪Y)/P(I);

The effective strong association rule analysis unit 20 is configured to determine a valid strong association rule according to the dependency relationship between the target frequent item set calculation modules.

Optionally, the effective strong association rule analysis unit 20 may include: a first analysis subunit 201 configured to calculate a confidence level of each non-empty item set in the target frequent item set, and a non-pre-set confidence threshold The empty item set is pruned; the second analysis sub-unit 202 is configured to calculate a degree of lifting between the non-empty item sets according to the confidence of the non-empty item set left after the pruning; the third analysis sub-unit 203, The rule set to determine that a non-empty item set that satisfies a degree of lift greater than 1 is a valid strong association rule.

Optionally, the first analysis sub-unit 201 may be further configured to: when calculating the confidence of each non-empty item set in the target frequent item set, adopt the following formula:

Confidence(X->Y)=P(Y|X)=P(X∪Y)/P(X);

Optionally, the second analysis sub-unit 202 may be further configured to calculate a degree of lifting between the non-empty item sets according to the confidence of the non-empty item set left after the pruning, using the following formula:

Lift(X->Y)=P(Y|X)/P(Y)=P(X∪Y)/P(Y);

If Lift(X->Y)>1, it means that rule X->Y is a valid strong association rule; if Lift(X->Y)<=1, it means that rule X->Y is invalid strong association rule. If Lift(X->Y)=1, it means that item set X and item set Y are independent of each other.

The association compilation building unit 30 is arranged to bind the modules that are effectively strongly associated together to compile or simultaneously warn.

Optionally, the modules with effective strong associations are combined and compiled at the same time to maximize self-healing, and the modules with effective strong associations are combined and early warning can reduce the human resources of manual troubleshooting errors and improve effectiveness.

Optionally, the apparatus may further include a correction unit configured to acquire the item set of the failed build and add to the knowledge base each time the compilation fails.

It should be noted that the foregoing device embodiments are the same as the method embodiments, and the implementation process may refer to the method embodiments, and the technical features in the method embodiments are applicable in the device embodiments.

The version software compiling apparatus provided in this embodiment automatically calculates the association degree between the modules by constructing the information knowledge base according to the pre-established failure, and determines the effective strong association rule; and binds the effective strong association modules together according to the effective strong association rule. Compile or alert at the same time. It can not only self-repair the compiled project, but also ensure continuous integration and reduce the risk of version release. It can also save the wrong human resources when the compilation fails.

Through the description of the above embodiments, those skilled in the art can clearly understand that the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. The software may be stored in a storage medium, which may be a non-transitory storage medium, including: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (Random Access Memory, A medium that can store program codes, such as a RAM, a disk, or an optical disk, or a transitory storage medium. The software may include a plurality of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform any of the methods described in the above embodiments.

As shown in FIG. 7, a general hardware structure diagram of a server provided in this embodiment is shown in FIG. 7. The server includes: a processor 310 and a memory 320; and may also include a communication interface (Communications). Interface) 330 and bus 340.

The processor 310, the memory 320, and the communication interface 330 can complete communication with each other through the bus 340. Communication interface 330 can be used for information transmission. The processor 310 can call the logic instructions in the memory 320 to perform the version software compilation method of the above embodiment.

In addition, the logic instructions in the memory 320 described above may be implemented in the form of a software functional unit and sold or used as a stand-alone product, and may be stored in a computer readable storage medium.

Finally, it should be understood that those skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by executing related hardware by a computer program, and the program can be stored in a non-transitory computer. In reading a storage medium, the program, when executed, may include a flow of an embodiment of the method described above, wherein the computer readable storage medium may be a magnetic disk, an optical disk, a read only memory (ROM), or a random access memory. (RAM), etc.

Industrial applicability

The present disclosure provides a version software compiling method and device, which can implement self-repair of a compiled project, ensure continuous integration effect, and reduce the risk of version release. It is also possible to save the wrong human resources when the compilation fails.

Claims

A version software compilation method, including:

Determining a target frequent item set in the knowledge base according to the pre-established failure building information knowledge base, wherein the failed build information knowledge base stores the item set that was previously failed to be built;

Determining a valid strong association rule according to the dependency relationship between the target frequent item set calculation modules;

Modules that are effectively strongly associated are bound together for simultaneous compilation or simultaneous warning.
The method of claim 1 further comprising:

When each compilation fails, the set of items that failed to build is acquired and added to the knowledge base.
The method according to claim 1, wherein the determining the target frequent item set in the knowledge base according to the pre-established failure building information knowledge base comprises:

Calculating a frequent item set of each order according to the item set of the knowledge base;

Calculating the support degree of each first-order item set in the knowledge base, and pruning the first-order item set whose support degree is lower than the first threshold, to obtain a first-order frequent item set;

According to the m-1 order frequent itemsets, multiple m-th order frequent itemsets are generated into multiple m-th order itemsets by lexicographic order and combination, and the support degrees of multiple m-th order itemsets are calculated respectively. The m-th order item set lower than the preset support degree threshold corresponding to the preset mth-th order item set is pruned to obtain an m-th order frequent item set;

When the number of m-th order frequent itemsets is one, it is determined that the m-th order frequent itemsets are target frequent itemsets, and the m+1th order item set is stopped; wherein m is a positive integer and m is greater than 1.
The method of claim 3, wherein calculating the support of the plurality of m-th order items, respectively, comprises:

Calculate the support of multiple m-th order items by using the following formula:

Support(X)=P(X)/P(I);

Support(X->Y)=P(X∪Y)/P(I);

Where I represents the total item set in the knowledge base, Support(X) represents the probability of the item set {X} appearing in the total item set, and Support(X->Y) indicates that the item set {X, Y} is in the total item. The probability of occurrence in the set, P represents the probability, P(X)/P(I) represents the probability that the item set (X) appears in the total item set, and the support degree of the item set refers to the probability that the item set appears in the total item set.
The method according to any one of claims 1 to 4, wherein the determining the effective strong association rule according to the dependency relationship between the target frequent item set calculation modules comprises:

Calculating a confidence level of each non-empty item set in the target frequent item set, and pruning the non-empty item set corresponding to the confidence level of the preset reliability threshold to obtain a plurality of target non-empty item sets;

Calculating the degree of elevation of each of the plurality of target non-empty item sets according to the confidence of the plurality of target non-empty item sets;

The rule between determining multiple target non-empty item sets with a degree of lift greater than 1 is a valid strong association rule.
The method according to claim 5, wherein said calculating a confidence level of each non-empty item set in said target frequent item set uses the following formula:

Confidence(X->Y)=P(Y|X)=P(X∪Y)/P(X);

Among them, the confidence Confidence (X->Y) indicates that in the case of having the item set X, the probability of having the item set Y is derived by the association rule "X->Y".
The method according to claim 6, wherein said calculating a degree of lifting between the plurality of target non-empty item sets according to the confidence of the plurality of target non-empty item sets, using the following formula:

Lift(X->Y)=P(Y|X)/P(Y)=P(X∪Y)/P(Y);

Wherein, the lift Lift(X->Y) represents the ratio of the probability of containing the item set Y under the condition that the item set X is included, and the probability of containing the item set Y under the condition that the item set X is not included;

If Lift(X->Y)>1, it means that the rule X->Y is a valid strong association rule;

If Lift(X->Y)<=1, it means that the rule X->Y is an invalid strong association rule;

If Lift(X->Y)=1, it means that item set X and item set Y are independent of each other.
A version software compiling device comprising:

a target frequent item set determining unit, configured to determine a target frequent item set in the knowledge base according to a pre-established failure building information knowledge base, wherein the failed build information knowledge base stores an item set that is previously failed to be built ;

An effective strong association rule analysis unit, configured to calculate between the modules according to the target frequent item set Dependencies to determine effective strong association rules;

Associate the compilation building unit, set to bind the modules with valid strong associations together and compile at the same time.
The apparatus of claim 8, wherein the target frequent item set determining unit comprises:

a first determining subunit, configured to sequentially calculate a frequent item set of each order according to the item set of the knowledge base;

a second determining subunit, configured to calculate a support degree of each 1st order item set in the knowledge base, and perform pruning on a 1st order item set whose support degree is lower than the first threshold to obtain a 1st order frequent item set;

The third determining subunit is configured to generate a plurality of m-th order item sets of the plurality of m-1 order frequent itemsets according to the m-1 order frequent item set, using a dictionary order and a combination manner, and respectively calculate a plurality of m-th order item sets The degree of support is prune to the m-th order item set whose support degree is lower than the preset support degree threshold corresponding to the m-th order item set, and the m-th order frequent item set is obtained. When the number of m-th order frequent itemsets is one, it is determined. The m-th order frequent item set is a target frequent item set, and the calculation of the m+1th order item set is stopped, where m is a positive integer and m is greater than 1.
The apparatus according to claim 8 or 9, wherein the effective strong association rule analysis unit comprises:

a first analysis subunit, configured to calculate a confidence level of each non-empty item set in the target frequent item set, and pruning a non-empty item set lower than a preset reliability threshold to obtain a plurality of target non-empty item sets ;

a second analysis subunit, configured to calculate a degree of elevation between the plurality of target non-empty item sets according to a confidence level of the plurality of target non-empty item sets;

The third analysis sub-unit is configured to determine that the rule between the plurality of target non-empty item sets having a degree of lift greater than 1 is a valid strong association rule.
A computer readable storage medium storing computer executable instructions for performing the version software compilation method of any of claims 1-7.