KR20140125949A - Sequence diagram generating method and system - Google Patents

Abstract

The present invention relates to a method for generating a sequence diagram and a system thereof. More specifically, the method includes: a step where an input unit receives input of source code and execution-tracing log data of software for maintenance; a step where a preprocessor removes unnecessary objects from the source code and execution-tracing log data; a step where a call tree generation unit detects a method call relation among nodes from the source code and execution-tracing log data and generates a static call tree and a dynamic call tree based on the detected call relation; a step where a pruning processing unit removes at least a node which has method data unnecessary for a functional flow in the generated static and dynamic call trees; a step where a call tree integration unit integrates the static and dynamic call trees from which the node having the unnecessary method data is removed into one call tree; and a step where a sequence diagram generation unit converts the integrated call tree to generate a sequence diagram. Accordingly, by generating a sequence diagram presenting the functional flow of software to be maintained, the method for generating a sequence diagram and the system thereof given in the present invention are able to execute the software quickly and easily through analysis of the generated sequence diagram.

Description

[0001] The present invention relates to a method and system for generating a sequence diagram,

The present invention relates to a method and system for generating a sequence diagram, and more particularly, to a method and system for generating a sequence diagram of software to perform maintenance, will be.

Due to the rapid development of IT technology, the scale of the software has increased and the complexity has increased gradually. This increases the maintenance cost of the software, which typically accounts for about 60% of the total software development cost during the maintenance phase of the entire life cycle of the software. Therefore, from a cost point of view, in order to reduce the development cost of the entire software, a method for reducing the maintenance cost must be performed.

In order to perform software maintenance, it is necessary to understand the existing system. This is because, in order to perform the maintenance work, the developer must accurately understand the effect of the change of the specific function on the entire system in the process of changing or expanding the existing system.

Nevertheless, understanding and reviewing the system at the maintenance stage is not performed smoothly due to lack of understanding of modeling and lack of modeling skill in the actual software development site. Accordingly, due to the absence of the software design model, it took about 30% of the maintenance time of the entire system to grasp the system to be maintained.

As described above, a method and system for generating a sequence diagram will be described as follows. Prior Art 1 is Korean Patent Publication No. 2004-0024923 (Mar. 24, 2004), which relates to a method and system for measuring software maintenanceability. This prior art 1 is a method and system for measuring or predicting software maintenanceability based on a procedure-oriented, object-oriented, component-based development method, or an output based on an advanced programming language, An object and environment variable input unit for automatically measuring maintenance of the software; A software requirements specification analysis engine that analyzes the software requirements specification and extracts measurement information; A software maintainability metric database that plans, creates, manages, and provides software maintainability metrics; A software maintainability metric setting engine that uses the software maintainability metric database and sets a software maintainability metric based on information extracted from the software requirements specification analysis engine; A software maintenance predictive model base that plans, generates, manages, and provides statistical or numerical software maintenance predictive models that are used to predict the maintainability of the software; Software scale and complexity measurement result database, software function measurement result database, software reliability measurement result software maintenance information by applying the information provided in the database and measuring the maintainability of the software by the set criteria, the selected prediction model and the extracted measurement information Measuring engine; And a measurement result output unit for outputting measurement results, thereby automatically and conveniently measuring the maintainability of the software in a format desired by the user in various software application fields, development methods, and programming language environments, thereby minimizing the time and economic loss . The present invention also provides a systematic and scientific approach to business management and engineering by providing convenient and accurate software maintenance measurement results. Further, the use or application of the invention is enterprise management, engineering and project management and is effective in measuring or predicting productivity, quality, cost, duration, risk, manpower and resources.

Prior Art 2 is Korean Patent Publication No. 2003-0018720 (Mar. 2003), which relates to a method and system for automatically generating a source code browser for software maintenance. The prior art 2 is a method for generating a source code browser for software maintenance, comprising: a first step of performing compilation and translation on a source code and extracting file and function information necessary for generating the source code browser; A second step of parsing the source code; Extracting a keyword from the parsed source code and highlighting it; Extracting a comment from the parsed source code and highlighting it; A fifth step of generating a hyperlink between the global variable / macro / data structure of the source code and the information corresponding to the global variable / macro / data structure based on the extracted file and function information; And analyzing the block structure of the source code based on the starting point of the block input from the user, the method comprising the steps of: In addition to the highlighting function for the code, it is possible to provide more convenient hyper function by using the information of the user-defined variable, thereby improving the understanding of the user's program and easily distinguishing the start and end of the block In addition to increasing visibility in the maintenance of programs as they can search in real time, software metrics (total lines, comment lines, blank lines, execution lines, etc.) metric) to analyze the source code to be maintained The time, effort, and cost required for the maintenance work of the program can be minimized, and the source code browser having various forms can be automatically generated.

In order to solve the problems of the related art as described above, the present invention can quickly and easily analyze the function flow of the software through the generated sequence diagram by generating a sequence diagram for the software to be maintained And to provide a method and system for generating a sequence diagram.

According to an aspect of the present invention, there is provided a method of generating a sequence diagram, the method comprising: receiving input source code and execution trace log information for software to perform maintenance or repair; Removing unnecessary objects from the source code and execution trace log information received by the preprocessor; The call tree generating unit detects a method call relationship between nodes from the source code and execution trace log information, and generates a static call tree and a dynamic call tree based on the detected call relationship; Removing at least one node having the static call tree in which the pruning processing section is created and the method information unnecessary in the function flow in the dynamic call tree; Integrating the static call tree and dynamic call tree into which the call tree integration side node has been removed into one call tree; And converting the one call tree in which the sequence diagram generating unit is integrated to generate the sequence diagram.

In particular, the method may include generating a static call tree and a dynamic call tree, wherein the call tree generator for detecting a call relationship of the inter-node method obtained from the source code and generating a static call tree based on the detected call relationship have.

In particular, the step of generating a static call tree and a dynamic call tree by the call tree generating unit for detecting a call relationship of the inter-node method obtained from the execution trace log information and generating a dynamic call tree based on the detected call relationship .

In particular, the call tree generating unit for generating a dynamic call tree using log information for recording a start point and an end point of execution of the method with respect to inter-node methods obtained from the execution trace log information includes a static call tree and a dynamic call tree For example.

More preferably, the pruning processing unit for returning a predetermined property value, judging a method that is commonly used or repeatedly executed as an unnecessary method in a function flow, And removing at least one node having information.

More preferably, the sequence diagram generator for obtaining the call sequence information of the method from the structure of the integrated call tree and converting the method to the sequence diagram based on the call sequence information of the obtained method, And generating a sequence diagram by transforming the sequence diagram.

According to another aspect of the present invention, there is provided a system for generating a sequence diagram, the system including: an input unit for receiving source code and execution trace log information for software to be maintained; A preprocessor for removing unnecessary objects from the input source code and execution trace log information; A call tree generation unit that detects a method call relationship between nodes from the source code and execution trace log information, and generates a static call tree and a dynamic call tree based on the detected call relationship; A pruning processing unit for removing the generated static call tree and at least one node having unnecessary method information in the function flow in the dynamic call tree; A call tree integration unit for integrating the static call tree and the dynamic call tree in which nodes are removed into one call tree; And a sequence diagram generation unit for generating a sequence diagram by converting one integrated call tree.

In particular, it may include a structure of a call tree that includes at least one of a set of nodes that invoke a method including a single root node, a set of methods in which calls are made in one direction, and a set of methods in which iterative calls are made have.

In particular, it may include a node including at least one of a class name, a method name, a parameter of a return value, a set of instructions to be executed when a method is invoked, and a current call state.

In particular, it may include a call tree generation unit that detects a call relationship of the inter-node method obtained from the source code, and generates a static call tree based on the detected call relationship.

In particular, it may include a call tree generation unit that detects a call relationship of the inter-node method obtained from the execution trace log information, and generates a dynamic call tree based on the detected call relationship.

In particular, the method may include a call tree generation unit that generates a dynamic call tree using log information that records a start point and an end point of execution of the method with respect to inter-node methods obtained from the execution trace log information.

In particular, it may include a pruning processing unit that returns a predetermined property value, or uses a general-purpose or repeatedly-performed method as an unnecessary method in a function flow.

More preferably, it may include a sequence diagram generator for obtaining call sequence information of the method from the structure of the integrated call tree, and converting the method to the sequence diagram based on the call sequence information of the obtained method.

The method and system for generating the sequence diagram of the present invention can generate a sequence diagram indicating the flow of the function of the software to perform maintenance so as to analyze the generated sequence diagram and analyze the function flow of the software quickly and easily It is effective.

Also, the method and system for generating the sequence diagram of the present invention can quickly analyze the function flow of the software to perform maintenance through the analysis of the sequence diagram, reduce the analysis time of the software, The time and cost required can be reduced.

1 is a block diagram of a generation system of a sequence diagram according to an embodiment of the present invention.
2 is a flowchart of a method of generating a sequence diagram according to another embodiment of the present invention.
3 is a diagram showing an example of a static call tree structure.
4 is a diagram illustrating an integration process of a call tree by applying a second aggregation rule.
5 is a diagram illustrating an integration process of a call tree by applying a third integration rule.
6 is a diagram illustrating an integration process of a call tree by applying a fourth integration rule.
FIG. 7 is a diagram illustrating an integration process of a call tree by applying a fifth integration rule. FIG.
8 is a block diagram illustrating a tool that implements the generation of the sequence diagram of the present invention.
9 is a view showing execution trace log information and a message to be removed.
10 is a diagram showing a generated sequence diagram.
11 is a diagram showing a pruning processing algorithm of the pruning processing unit.
12 is a diagram illustrating an algorithm for the integration process of a call tree.
13 is a diagram showing system information of an object of the present invention.
14 is a diagram comparing the number of nodes in the call tree.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Hereinafter, the present invention will be described in detail with reference to preferred embodiments and accompanying drawings, which will be easily understood by those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein.

Hereinafter, a system for generating a sequence diagram according to an embodiment of the present invention will be described in detail with reference to FIG.

1 is a block diagram of a generation system of a sequence diagram according to an embodiment of the present invention.

1, the generation system 100 of the sequence diagram of the present invention includes an input unit 110, a preprocessing unit 120, a call tree generation unit 130, a pruning processing unit 140, (150) and a sequence diagram generator (160).

The input unit 110 receives the source code and execution trace log information for the software to be maintained.

The preprocessing unit 120 removes unnecessary objects from the input source code and execution trace log information.

The call tree generation unit 130 detects a method call relationship between nodes from the source code and execution trace log information, and generates a static call tree and a dynamic call tree based on the detected call relationship. Thus, the structure of the generated static or dynamic call tree is composed of a plurality of node sets that call a method including one root node, a set of methods in which calls are made in one direction, and a set of methods in which repeated calls are made As shown in FIG. At this time, the node that calls a method may include at least one of a name of a class, a name of a method, a parameter of a return value, a set of instructions to be executed when a method is called, and a current call state. The call tree generation unit 130 detects a calling relationship of the method obtained from the source code and generates a static call tree based on the detected calling relationship. At this time, the AST (Parser) generated by the Eclipse's AST parser (Abstract Syntax Tree) to generate a static call tree. Also, the call tree generation unit 130 detects a call relationship of the method obtained from the execution trace log information, and generates a dynamic call tree based on the detected call relationship. In particular, the call tree generating unit 130 generates a dynamic call tree by using log information that records a start point and an end point of execution of the method with respect to the called method obtained from the execution trace log information. At this time, Use the open source Btrace tool to create a dynamic call tree.

The pruning processing unit 140 removes the generated static call tree and nodes having unnecessary method information in the function flow in the dynamic call tree. The pruning processing unit 140 returns a predetermined property value, or widely used method, or determines a method to be performed as an unnecessary method in a function flow.

The call tree integration unit 150 integrates the static call tree and the dynamic call tree from which the node has been removed into one call tree.

The sequence diagram generation unit 160 generates a sequence diagram by transforming one integrated call tree. The sequence diagram generator 160 obtains the call sequence information of the method from the structure of the integrated call tree and converts the method into the sequence diagram based on the call sequence information of the obtained method.

Hereinafter, a method of generating a sequence diagram according to another embodiment of the present invention will be described in detail with reference to FIG.

2 is a flowchart of a method of generating a sequence diagram according to another embodiment of the present invention.

As shown in FIG. 2, in the method of generating the sequence diagram of the present invention, the input unit 110 receives source code and execution trace log information for software to be maintained (S210).

The preprocessing unit 120 removes unnecessary objects from the source code and the execution trace log information (S220).

Particularly, the conventional sequence diagram has a great difference in the complexity of the sequence diagram depending on the calling depth of the method, ie, how many steps are called from the first method to be called. In addition to the class implemented in the source code, , The use of the framework, and the like, all of the objects to be referred to include the complexity. In addition, the complexity of the sequence diagram becomes very large when all the contents existing in the source code are reflected. Therefore, in order to reduce the complexity of the sequence diagram to be generated later and to set the analysis range of the target software, the object to be referred to at the time of initial creation of the software is limited to the package designated by the user, and the level of the object represented in the sequence diagram is stored in the memory All objects created on the object are abstracted at the class level, which is an object type, and are expressed in a sequence diagram. In addition, by excluding Java standard libraries or commercial libraries that are not selected by the user, the level of nesting is also automatically limited.

The call tree generation unit 130 detects a method call relationship from the source code and the execution trace log information, and generates a static call tree and a dynamic call tree based on the detected method call relationship (S230). At this time, The generation unit 130 detects the calling relation of the method obtained from the source code, and generates the static calling tree based on the detected calling relation. Alternatively, the call tree generating unit 130 detects a call relation of a method obtained from the execution trace log information, and generates a dynamic call tree based on the detected call relationship.

In particular, the call tree generating unit 130 generates a dynamic call tree using log information that records the start and end points of execution of the method for the called method obtained from the execution trace log information.

That is, a flow in which a call of a method performing each function of the target software is performed can be expressed in the form of a call tree. The structure of such a call tree is a graphical representation of the calling relationship of a method, and is a kind of graph that expresses a relation of calling a method to a node as an edge. Unlike a normal graph, the structure of such a call tree has only one direction from a method to another method, and does not have a circular relationship between methods. In addition, the call tree structure has a root node indicating the start of a call for one method, so that all nodes have a level (step) indicating a distance from one root node. Thus, unlike a sequence diagram that focuses on the method invocation order, the call tree focuses on the relationship between the invoking method (caller) and the invoked method (callee) useful.

Hereinafter, the static call tree structure of the call tree structure will be described in detail with reference to FIG.

3 is a diagram showing an example of a static call tree structure.

As shown in FIG. 3, each node constituting the static call tree is represented by a rectangle. At this time, A.a represented in the node represents a method a belonging to the class A. In particular, since the AST used to create the static call tree does not know the class of the calling method, it obtains the class information of the calling method from the tree structure of Eclipse's project explorer. At this time, each node has a relation with the class and a signature of the method as an attribute. In particular, in order to indicate the reduction in the recursion of the method that appears repeatedly, the iterative method is defined as a 'iterative method block', displayed as a rectangle, and then displayed as a circular arrow. That is, as shown in FIG. 3, the node C.c and the node F.f are grouped into a single rectangle, and an arrow circulating in the rectangle is displayed to indicate that the node C.c and the node F.f are repetitive nodes.

In addition, if there is a method node appearing more than once in the static call tree, the representation of the duplicated child node of the corresponding method node is omitted, and * is additionally displayed in the corresponding method node. That is, as shown in FIG. 3, since the node Bb appears twice, the notation of the child node Dd and the node Ee of the node Bb is omitted at the time of the second appearance of the node Bb, do.

In addition, let's look at the mapping relation with the call tree according to the call relation acquired from the source code.

The pruning processing unit 140 removes the static call tree from which the static call tree is generated and the node having unnecessary method information in the function flow in the dynamic call tree at step S240. The pruning processing unit 140 may return a predetermined property value or determine a general-purpose method as an unnecessary method in the function flow, and send a node having the method information determined to be unnecessary as the static call tree or dynamic call Remove from the tree.

Hereinafter, the unnecessary methods will be described in detail.

For example, a getter method is a method that is designed to return a predefined attribute value, a method that does not affect the functionality of the system, except to access the attribute to read the value. Utility methods are not limited to specific functions but represent methods that can be used in various parts of the system. Many software systems use utilities to help implement the functions of the system, A routine or subsystem, or a separately implemented class.

In this manner, it is possible to determine that the above-described method is unnecessary for the function flow, and to remove the node having the information of the method from the static call tree or the dynamic call tree.

Therefore, by decreasing the number of messages included in the sequence diagram to be generated later, the size of the sequence diagram can be reduced, and unnecessary information can be removed, so that the function flow can be represented simply and clearly.

The call tree integration unit 150 integrates the static call tree and the dynamic call tree from which nodes having unnecessary method information are removed into one call tree at step S250.

The integration process of the call tree is performed based on the first to fifth integration rules set in advance.

First, the first aggregation rule identifies a common method node between the static call tree and the dynamic call tree.

That is, it is a dynamic call tree current method nodes of T _D M (i, d) and the static call tree T _S of the present method the node M (j, d) are the same when the method nodes integrated call tree T _C M (i , d) are added.

In addition, the second aggregation rule eliminates duplication of the message invocation sequence.

For example, if the current method node M (j, d) of the static call tree T _S corresponds to an iterative method block, it is searched whether there is an iteration of the same method at the dth depth of the dynamic call tree T _D , , The iterative method block of the static call tree T _S is added to the unified call tree T _C.

4 is a diagram illustrating an integration process of a call tree by applying a second aggregation rule.

As shown in Fig. 4 (a), there are repeated method blocks Cc, D, d in the static call tree T _S and a repetition sequence of the same methods in the dynamic call tree T _D (Cc → Dd → Cc → Dd ) Is present in the region. Accordingly, by integrating the static call tree T _S and the dynamic call tree T _D by applying the second integration rule, the integrated call tree T _C has the repeated method blocks Cc and Dd.

Specifically, FIG. 4 (b) then shows the sequence diagram generated from the dynamic call tree T _D , and FIG. 4 (c) shows the sequence diagram generated from the integrated call tree T _C afterwards. As shown in Fig. 4 (c), the repetitive method block is represented as a loop frame, making it possible to make it more compact as compared with the sequence diagram shown in Fig. 4 (b).

In addition, the third unified rule is to identify the objects that actually participate.

For example, if the current method node M (i, d) of the dynamic call tree T _{D and} the current method node M (j, d) of the static call tree T _S have the same signature, In other cases, the relationship between the two method nodes is retrieved. Then, the dynamic call tree T _D of the current method, the node M if (i, d) classes are available for the children of the class's static call tree T _S Current methods node M (j, d) as in the integrated call tree Tc Adds the current message node M (i, d) of the dynamic call tree T _D.

5 is a diagram illustrating an integration process of a call tree by applying a third integration rule.

Since the, was identified by the method d () in the node Ed node Dd and dynamic call tree T _D of the static call tree, class E, class D child class 5, the dynamic call tree T _D of the node Ed is added to the unified call tree T _C by the third unified rule. Accordingly, in the sequence diagram to be generated later, d () of the actually called class E appears as a message, and when compared with the sequence diagram generated from the dynamic call tree T _D , the sequence diagram generated from the integrated call tree T _C , You can specify a class that actually participates in the system so that you can understand system behavior more clearly.

In addition, the fourth union rule identifies unexecuted methods.

For example, current methods node of the static call tree Ts M (i, d) and a dynamic call tree T _D of the current method, the node M (j, d) for the cross-comparison, the dynamic call tree T _D the current method the node M of (j, d) is null, adds the current method node M (i, d) of the static call tree Ts to the unified call tree Tc.

Alternatively, if the current method node M (j, d) of the dynamic call tree T _D is not null and is different from the node M (i, d), M (i, d) And again with the next node M (j + 1, d) of the call tree T _D.

Accordingly, if the node M (j + 1, d) is different or null from the node M (i, d), the node M (i, d) is added to the integrated call tree Tc.

Then, the method node M (i, d) of the static call tree Ts is displayed as a message of the option frame in the generated sequence diagram.

6 is a diagram illustrating an integration process of a call tree by applying a fourth integration rule.

As shown in Figure 6, the method node Dd of the static call tree Ts does not exist in the dynamic call tree T _D because the d () method of class D is implemented but not executed when generating the execution trace log to be. As a result, the method node Dd is added to the integrated call tree T _C using the fourth integration rule, so that in the sequence diagram generated from the dynamic call tree T _D , the missing method d () is shown in the sequence diagram generated from Tc And can improve the coverage of reverse engineering.

Finally, the fifth unification rule is to identify the late binding behavior.

For example, the current method node M (j, d) of the dynamic call tree T _D is compared with the current method node M (i, d) of the static call tree Ts.

If the current method node M (i, d) of the static call tree Ts is null, a new static call tree Ts' having the current method node M (j, d) of the dynamic call tree T _D as the root node is generated from the source code , And adds the root node of the generated new static call tree Ts' to the next neighbor node of the M (i, d) node of the previous static call tree Ts.

If the M (i, d) is not null and is different from M (j, d), the current method node M (j, d) of the dynamic call tree T _D and the next neighboring node M (i + ).

If the next neighboring node M (i + 1, d) is a different node from the current method node M (j, d) or null, a new static call tree Ts' Add the root node to the next neighbor node of the current method node M (i, d) of the previous static call tree Ts.

Then, the comparison is restarted from the current method node M (j, d) of the dynamic call tree T _{D and} the current method node M (i, d) of the static call tree Ts.

FIG. 7 is a diagram illustrating an integration process of a call tree by applying a fifth integration rule. FIG.

As shown in Fig. 7, the method d () of the node Dd of the dynamic call tree T _D can know that the method d () is dynamically bound through what is not identified in the static call tree Ts. The sequence diagram generated from the static call tree Ts consists of only three method calls Aa, Bb, and Cc. The sequence diagram generated from the integrated call tree Tc generated by applying the fifth integration rule described above includes not only the method Dd , And other method calls Ee, Ff, and Gg. Particularly, the first integrated rule and the fourth integrated rule are also applied to the integrated call tree Tc through the continuous comparison process between the static call tree and the dynamic call tree. As a result, the final integrated call tree as shown in FIG. 10 (c) have.

Thus, it can be seen that the generated sequence diagram is solved based on the call tree in which the static call tree and the dynamic call tree are integrated, thereby solving the problem of disconnecting the function flow of the static call tree. In addition, it can be seen that the problem of the missing function flow of the dynamic call tree is solved by reflecting the system context information based on the static call tree.

Hereinafter, an actual implementation of the sequence diagram according to the present invention will be described in detail with reference to FIG. 8 through FIG.

First, we developed the Rich Client Platform (RCP) in the Eclipse project. The Rich Client is a client that processes a large amount of data on its own, and does not depend on the server. Also, by using Graphical Editing Framework, information of integrated call tree is generated in the form of sequence diagram.

The Rich Client Platform (RCP), which is a component of the system architecture of the RCP that is the basis of this tool, and the component architecture of the tool, is Java-based platform-independent, provides a basic user interface and supports Graphical Editing Framework (GEF) Is an Eclipse framework that makes it easy to create high-quality, time-based editors and views, and provides integration with the Eclipse workbench user interface.

8 is a block diagram showing a tool that implements the generation of the sequence diagram of the present invention.

As shown in Fig. 8, the log reader converts the call relationship log extracted from the source code into analysis data through execution trace log and Java parser.

The Call Tree Builder generates the analysis data in the form of a call tree based on source code and a dynamic tree based on execution trace.

Call tree pruner removes unnecessary method information and its subflows from the two generated call trees.

The call tree combiner combines the two generated call trees into one call tree according to the rules.

Sequence diagram The combiner reads and translates the integrated call tree into a sequence diagram.

The Model Editor reads the information of the created sequence diagram and displays it on the screen.

The Project Manager manages a single project unit for a target system that contains multiple sequence diagrams.

The user interface (UI) portion of the functions of the above-described sequence diagram tool will be described in detail.

9 is a view showing execution trace log information and a message to be removed.

As shown in FIG. 9, the first part is a part showing a project managing the generated sequence diagram. The second part is a log containing the name of the sequence diagram, a log containing the call relationship information extracted from the source code by using Java Parser, a log containing information of the method executed by the system using the Java Profiling tool . Finally, section 3 shows information of all the currently included methods based on the input log. Among the displayed methods, the user can select the method to be removed, and the selected information can be saved as a separate filter file for later use.

Hereinafter, a generation screen of the sequence diagram will be described with reference to FIG.

10 is a diagram showing a generated sequence diagram.

As shown in FIG. 10, a sequence diagram is generated after all the log files are input and a method to be removed is selected. 1 shows the generated sequence diagram. 2 shows briefly the objects included in the corresponding sequence diagram, 3 shows the entire sequence diagram small, and 4 shows the attribute information of objects and messages in the sequence diagram. The generated sequence diagram is included in the project.

Before merging the two call trees, remove the node containing the method information you selected.

Hereinafter, an algorithm for removing the node will be described with reference to FIG.

11 is a diagram showing a pruning processing algorithm of the pruning processing unit.

As shown in FIG. 11, the algorithm for removing the corresponding node is expressed by Equation 1.

The method selected by the user is stored as data of type filter (1). (2) if all the methods stored in the filter are the same as the method information of the current node, and if they are the same, remove the node from the tree (3-5). If the current node is not the same as the method information stored in the filter, the search goes down to the current node's child nodes, and the method is recursively called (10).

Hereinafter, an algorithm for the integration process of the call tree will be described with reference to FIG.

12 is a diagram illustrating an algorithm for the integration process of a call tree.

As shown in FIG. 12, the two call trees refined through the pruning process are integrated into one call tree.

First, the child nodes of the node selected in the static call tree and the dynamic call tree are stored in the array of node types of S and D (1), and the two arrays are integrated into one array. At this time, the merging is performed by adding nodes of D based on the nodes of S. Thus, the integration process is repeated until all the nodes of D are merged into S in order (3). First, the node matching the node of D is found in S and the index of that node is stored in lastIndex (6). lastIndex indicates the point where the integration is already completed. If D [1] matches S [3], then lastIndex is 3, which means that S [0] through S [2] already consider the integration of D [2] It means you do not need to look at it. (5) If the node corresponding to D [id] is in S (5), it updates the corresponding lastIndex and invokes the integration method recursively to proceed with the integration of the child nodes of the nodes (6).

If there is no matching node, it is checked whether a node having an overriding relation exists (9). If a node with an overriding relationship has a repeating statement, then it is preceded by a corresponding node (12 to 14) for the next merging process, otherwise the corresponding node is replaced by D [id] (16 ).

If there is no corresponding node, add D [id] to the end of S (19-20).

The merging process goes down to the child nodes of the current node, and the corresponding method is recursively called (22).

In order to evaluate the utility of the present invention, a case study was conducted on REQ_Modeler, which is an open source modeling program, TerpPaint, a painting program, and Spec Generator, a specification and a source code automatic generation tool for research, And the following.

13 is a diagram showing system information of an object of the present invention.

In exploiting the sequence diagram generated by the present invention, the main problem that arises is the incompleteness of flow due to the use of dynamic calls and external frameworks and external libraries, and the size of the sequence diagram. For this reason, the experimental subjects chose to utilize mainly the system or the external library which processes the user interface (UI) in which the dynamic call frequently occurs. We have selected the developed Java based analysis tool from Reverse Sequence Diagram Tool which is implemented for the experiment. In addition, in order to secure the objectivity of the experiment, REQ_Modeler and TerpPaint were selected in different categories among the open source programs in which the user manual and the source code are disclosed together as a basis for extracting functions among the programs distributed in the executable form .

For the experiment, we compare and analyze the representative functions in the target system with the static analysis tool Together, the dynamic analysis tool TPTP, and the sequence diagram generated by the proposed method.

14 is a diagram comparing the number of nodes in the call tree.

As shown in Fig. 14, the sizes of the generated static call tree, the dynamic call tree, and the integration tree can be determined according to the present invention with respect to the selection functions.

For the experiment, REQ_Modeler extracts 16 functions based on the usage manual. Among them, one representative function for similar functions, namely, class creation, use case generation, actor creation, And 8 representative functions (1 to 7) were extracted with the class generation function as a representative function. In the same way, five functions (8 ~ 12) were extracted in TerpPaint and two functions (13 ~ 14) were extracted in Spec Generator. The columns show the size of the static call tree, and the column shows the size of the dynamic call tree. The column ③ is the size of the tree which integrates the call tree of ① and ② according to the rule. The comparison of these tree sizes can be roughly classified into three kinds.

First, when the flow of the static call tree is disconnected, there is a dynamic call flow that can not be understood by the source code alone in execution of the function. REQ_Modeler and TerpPaint show that the user-interface (UI) -related logic consists of dynamic calls and that most of the functionality goes through the logic, so these figures are prominent.

Second, when the flow of the dynamic call tree is missing, there is a flow by various system contexts that were not recognized only by execution trace. For example, this happens in REQ_Modeler class move, delete, connection, etc. (4,5,6) because when a function is executed, only certain flows among functional flows are executed according to the state of the class.

Third, when a duplicate method of a dynamic call tree is included, iteration of certain logic occurs due to loops and so on. In the SpecGenerator (13, 14), which repeatedly reads and transforms specific information, dragging (8, 9) of TerpPaint which continuously calls a specific method to continue showing the operation while dragging the mouse, In the case of a paint bucket 11 of TerpPaint that calls methods as many times as there are corresponding pixels to fill a certain color.

Among the above three cases, the first case is mainly based on static analysis, the second and third cases are problems on dynamic analysis basis, and it has been pointed out in a lot of studies, and the integrated tree according to the present invention It can be seen that three types of problems are solved.

In addition, the method and system for generating such a sequence diagram may be stored in a computer-readable recording medium on which a program for execution by a computer is recorded. At this time, the computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer readable recording medium include ROM, RAM, CD-ROM, DVD 占 ROM, DVD-RAM, magnetic tape, floppy disk, hard disk, optical data storage, and the like. In addition, the computer-readable recording medium may be distributed to network-connected computer devices so that computer-readable codes can be stored and executed in a distributed manner.

The method and system for generating the sequence diagram of the present invention can generate a sequence diagram indicating the flow of the function of the software to perform maintenance so as to analyze the generated sequence diagram and analyze the function flow of the software quickly and easily It is effective.

Also, the method and system for generating the sequence diagram of the present invention can quickly analyze the function flow of the software to perform maintenance through the analysis of the sequence diagram, reduce the analysis time of the software, The time and cost required can be reduced.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, Do.

110: input unit 120: preprocessing unit
130: call tree generation unit 140: pruning processing unit
150: call tree integration unit 160: sequence diagram generation unit

Claims (15)

Receiving input source code and execution trace log information for software whose input unit is to perform maintenance;
Removing unnecessary objects from the source code and execution trace log information received by the preprocessor;
The call tree generating unit detects a method call relationship between nodes from the source code and execution trace log information, and generates a static call tree and a dynamic call tree based on the detected call relationship;
Removing at least one node having the static call tree in which the pruning processing section is created and the method information unnecessary in the function flow in the dynamic call tree;
Integrating the static call tree and the dynamic call tree into which a node having unnecessary method information has been removed, into a call tree; And
Generating a sequence diagram by transforming one call tree into which the sequence diagram generating unit is integrated;
≪ / RTI >
The method according to claim 1,
The step of the call tree generating unit generating the static call tree and the dynamic call tree
Detecting a call relationship of the inter-node method obtained from the source code, and generating a static call tree based on the detected call relationship.
The method according to claim 1,
The step of the call tree generating unit generating the static call tree and the dynamic call tree
Detecting a call relationship of the inter-node method obtained from the execution trace log information, and generating a dynamic call tree based on the detected call relationship.
The method of claim 3,
The step of the call tree generating unit generating the static call tree and the dynamic call tree
Wherein the dynamic call tree is generated by using log information for recording a start point and an end point of execution of the method with respect to inter-node methods obtained from the execution trace log information.
The method according to claim 1,
Wherein the pruning processing unit removes at least one node having unnecessary method information in the static call tree and the function flow in the dynamic call tree
A method for returning a predetermined property value, a method used universally, or a method repeatedly performed is determined as an unnecessary method in a function flow.
The method according to claim 1,
Wherein the step of generating a sequence diagram by converting one call tree integrated with the sequence diagram generating unit
Acquiring call order information of a method from a structure of an integrated call tree, and converting the method based on call order information of the obtained method to generate a sequence diagram.
A computer-readable recording medium on which a program for executing the method according to any one of claims 1 to 6 is recorded.
An input unit for receiving source code and execution trace log information for software to be maintained;
A preprocessor for removing unnecessary objects from the input source code and execution trace log information;
A call tree generation unit that detects a method call relationship between nodes from the source code and execution trace log information, and generates a static call tree and a dynamic call tree based on the detected call relationship;
A pruning processing unit for removing the generated static call tree and at least one node having unnecessary method information in the function flow in the dynamic call tree;
A call tree integration unit for integrating the static call tree and the dynamic call tree in which nodes are removed into one call tree; And
A sequence diagram generation unit for generating a sequence diagram by converting an integrated call tree;
/ RTI > of the sequence diagram.
9. The method of claim 8,
The structure of the call tree is
Wherein the at least one node set includes at least one of a plurality of node sets including a root node to call a method, a set of methods in which calls are made in one direction, and a set of methods in which repeated calls are made. Gt;
10. The method of claim 9,
The node
A name of a class, a method name, a parameter of a return value, a set of instructions to be executed when a method is called, and a current call state.
9. The method of claim 8,
The call tree generating unit
Detecting a call relationship of the inter-node method obtained from the source code, and generating a static call tree based on the detected call relationship.
9. The method of claim 8,
The call tree generating unit
Detecting a call relationship of the inter-node method obtained from the execution trace log information, and generating a dynamic call tree based on the detected call relationship.
17. The method of claim 16,
The call tree generating unit
Wherein the dynamic call tree is generated using log information that records a start point and an end point of execution of the method with respect to inter-node methods obtained from the execution trace log information.
9. The method of claim 8,
The pruning processor
A method for returning a predetermined property value, a method used universally, or a method repeatedly determined as an unnecessary method in a function flow.
12. The method of claim 11,
The sequence diagram generation unit
Acquiring call order information of a method from a structure of an integrated call tree, and converting the method based on call order information of the obtained method to generate a sequence diagram.

Images