JP5123350B2 - Test case creation system, method and program - Google Patents

Description

The present invention is a test case creation system to automatically create a test case from the document data specifications or the like, and relates to a method and program.

  During the development of a product such as software, the function and performance are tested. Such tests not only confirm that the product operates correctly as described in the specifications, but also detect product defects and defects, etc. to ensure product quality and reliability. In order to improve the performance of the test, use the words described in the specification or read the meaning of the sentence to create various test cases that are not directly described in the specification, and test more extensively and comprehensively. It is common to do.

  For example, when testing a program, it is necessary to check whether the various processes related to the input / output are performed correctly, but there are a huge number of patterns. Therefore, there is a problem that it takes a lot of time and labor to create a necessary test case from the items described in the specification. Also, in the past, test engineers often created test cases from specifications based on their own experience and judgment, so there was a problem that the selection was inevitably biased and it was difficult to ensure completeness. there were.

  In order to solve such problems, for example, as shown in Patent Document 1 and Patent Document 2, development of a system that automatically generates a test case from a specification has been attempted. However, even with these systems, although test cases can be generated from tables and state transition diagrams included in specifications, new test cases can be generated from sentences that occupy most of the specifications. Therefore, it still has a problem in terms of completeness. In addition, in the orthogonal table, when generating test cases by combining factors, there is an advantage that experiments on many factors can be performed in a relatively small number of times, but it is necessary for the test engineer to select the factors to be combined themselves. In many cases, unnecessary test cases are generated.

Japanese Patent Application Laid-Open No. 11-306044 JP 7-253905 A

The present invention has been made in view of such circumstances, and can significantly reduce the time and labor required to create a test case, and can prevent bias in test case selection. Accordingly, an object of the invention to provide a test case creation system, method and program that can realize a more comprehensive and efficient testing.

  A test case creation system according to the present invention includes a storage unit that stores character string data of a document in which product specifications are described, and reads each of the character string data from the storage unit, and each sentence included in the document A morpheme analysis execution means for storing the words in the first database in association with the part of speech determined by the morpheme analysis, using the minimum unit character or character string obtained by the morpheme analysis as a word, A word or combination of words satisfying a preset part-of-speech condition is acquired from the first database, and the acquired word or combination of words is replaced with another word or combination of words in a sentence including them. A test case generating means for generating a new sentence and storing it as a test case in the second database; It is characterized in that an output means for obtaining and outputting a test case from the database.

Here, the test case describes the items (preconditions, operations, and results that result) necessary for the test, and is used to verify whether products such as programs are operating correctly. It is.
Examples of words that satisfy a predetermined part-of-speech part of speech include, for example, the word “noun” (claim 3) registered in the category dictionary, and “adjective-independent” following “particle-case particle” “”. A word indicating negation (such as “no”, “prefix-noun connection” or “non”) immediately before “noun-sa-variant connection”, or “by” or “in” Examples include words that specify an object (claim 11), such as “particle-case particle-collocation”, and on the other hand, as a combination of words that satisfy a condition related to a preset part-of-speech, the viewpoints are the same and common Combinations of words including nouns or verbs (Claim 2) are included. “Negation” means “prefix-noun connection” (“not”, “non”, “non”, etc.), “auxiliary verb” and “adjective-independence” (no “no”) "," ~ "And" z "etc.).

  In the test case creation system, on the basis of the positional relationship of parts of speech determined by the morphological analysis, each sentence included in the document is excluded from a main prescription sentence that combines a subject clause and a predicate clause, and the main prescription sentence. A phrase classification means for classifying the information into one or a plurality of conditional clauses and storing the information in the first database in association with each word constituting the clauses, and the test function, condition, operation, execution, and results When the viewpoints are the functional viewpoint, the conditional viewpoint, the operational viewpoint, the execution viewpoint, and the results viewpoint, respectively, and the words that are the factors of the respective viewpoints are the function factor, condition factor, operation factor, execution factor, and result factor, For each of these factors, the words stored in the first database are identified as part of speech determined by the morphological analysis and the position in the sentence. And classifying means that associates the classification result with each word and stores it in the first database, and in the case where the same factor continues in a sentence, If not, viewpoint generating means for generating individual factors as functional viewpoints, condition viewpoints, operation viewpoints, execution viewpoints or results viewpoints, and storing these viewpoints in association with the respective words in the first database, The test case generation means takes out combinations of words classified as the same kind of viewpoint from the first database, extracts a combination of words including common nouns or verbs, and then extracts them. Generate new sentences by substituting each other in a sentence that contains word combinations, except for word combinations that completely match the word combinations. It may be one stored in the test case generation DB as the test case.

  Here, the functional viewpoint, condition viewpoint, operation viewpoint, execution viewpoint, and result viewpoint are the combinations determined by the part of speech required in the processing of “5.3: Test case generation step (replacement of word combinations)” described later. A word or a single word.

  In addition, the test case creation system may have a category dictionary in which words that are nouns are registered by category. In that case, the test case generation unit extracts words that become nouns from the database. Thereafter, it is determined whether or not the word is present in the category dictionary, and when it is determined that the word is present in the category dictionary, another word belonging to the same category as the word is acquired from the category dictionary. Then, a new sentence may be generated by replacing the word with the other word in a sentence including the word, and stored as a test case in the second database.

  Further, the test case creation system determines a dependency relationship between the sentences included in the document, and when the main sentence corresponds to the test case as a result of the determination, the dependent sentence is also used as the test case. While storing in the second database, if the main sentence does not correspond to the test case, the dependent sentence may be excluded from the test case.

  Specifically, as a condition for determining the dependency relationship between the sentences, for example, when the nouns in the sentence are extracted and the words of the same noun are included in the consecutive sentences, the previous sentence Is determined as a subordinate sentence (condition 1), and a sentence starting with a conjunction is determined as a sentence subordinate to the immediately preceding sentence (condition 2), “noun-pronoun-general” Sentences that start with or contain the words "noun-non-independent-adverb", "adverb-general" ("from these", "from (more than)", "for example", etc.)) Judge as a subordinate sentence (Condition 3). Extract nouns contained in sentences beginning with “From these”, “From above”, etc., and search for sentences using those nouns from the top of the document data. , If there is an appropriate sentence, that sentence is the main sentence, These "," and the like is determined respectively (condition 4) as a statement that a subordinate beginning statements like the above ".

  The test case generation means extracts a word representing negation from the first database based on the part of speech, and then replaces the word representing negation with a word representing affirmation in a sentence including the word. Thus, a new sentence may be generated and stored in the second database as a test case.

  In that case, the test case generation means affirms an affirmative sentence when a sentence including the word representing the negation and a new sentence generated from the sentence is a negative sentence. In the case of a form sentence, new sentences may be further generated by converting to a negative form sentence and stored in the second database as a test case.

  In addition, the test case generation unit extracts a new sentence by extracting a word specifying the target from the first database based on the part of speech, and then replacing the target of the sentence including the word with another target. May be generated and stored in the second database as a test case.

  Furthermore, the test case creation system may further include data extraction means for extracting character string data from the electronic data of the document and storing it in the storage means.

  The test case creation method according to the present invention includes storage means for storing character string data of a document in which product specifications are described, and control for generating a test case from the character string data stored in the storage means. A test case creation method executed by a computer comprising a control means for performing the morphological analysis of each sentence included in the document by reading the character string data from the storage means Storing a word or a character string of a minimum unit obtained by the morpheme analysis as a word, storing the word in the first database in association with the part of speech determined by the morpheme analysis, and the control means including the first database To obtain a word or combination of words that satisfies a pre-set part of speech condition, and A step of generating a new sentence by replacing it with another word or combination of words within the sentence in which they are included, and storing the new sentence as a test case in the second database; and the control means includes the second database And a step of obtaining and outputting a test case from the test case.

  A program according to the present invention includes a storage unit that stores character string data of a document in which product specifications are described, and a control unit that performs control to generate a test case from the character string data stored in the storage unit The control means reads the character string data from the storage means, performs morphological analysis of each sentence included in the document, and obtains the morphological analysis by the morphological analysis. And storing the words in the first database in association with the part of speech determined by the morphological analysis, and the control means is preset from the first database. Acquire a word or combination of words that satisfies the part-of-speech condition, and the acquired word or combination of words is a sentence that contains them. And generating a new sentence by replacing it with another word or combination of words and storing it as a test case in the second database; and the control means obtains the test case from the second database. And the step of outputting in a computer.

  According to the present invention, a morpheme analysis is performed using character string data of a document in which product specifications are described, and a word or a word satisfying a preset part-of-speech condition based on the result of the morpheme analysis. Acquire a combination, and replace the acquired word or combination of words with another word or combination of words in the sentence that contains them to generate a new sentence and use it as a test case As a result, the time and effort required to create test cases can be greatly reduced, and bias in test case selection can be prevented, resulting in more comprehensive and efficient testing. It becomes possible to do.

FIG. 1 is a block diagram showing an embodiment of a test case creation system according to the present invention. FIG. 2 is a block diagram showing a schematic configuration of the test case creation system. FIG. 3 is a flowchart showing the overall flow of the test case creation process. FIG. 4 is a diagram illustrating an example of electronic data of a document. FIG. 5 is a diagram illustrating an example of character string data after text conversion processing. FIG. 6 is a chart showing an example of the result of morphological analysis. FIG. 7 is a diagram for explaining a phrase classification method. FIG. 8 is a diagram for explaining a method of classifying main prescribed sentences. FIG. 9 is a diagram illustrating an example sentence 1-3 for explanation. FIG. 10 is a chart showing an example of factor classification results. FIG. 11 is a chart showing the continuation of the classification result of FIG. FIG. 12 is a diagram illustrating an example of the factor / viewpoint DB. FIG. 13 is a chart showing an example of viewpoint classification results. FIG. 14 is a flowchart illustrating a sentence dependency setting method. FIG. 15 is a diagram illustrating an example of a category dictionary. FIG. 16 is a flowchart showing a noun word replacement method. FIG. 17 is a chart illustrating an example of a word representing negation. FIG. 18 is a diagram illustrating an example of a negative / affirmative DB. FIG. 19 is a diagram illustrating a method of adding a test case by replacing affirmative / negative words. FIG. 20 is a diagram illustrating an example of a correspondence relationship between positive and negative forms used at the end of a sentence. FIG. 21 is a diagram illustrating an example of a test case generated from a sentence described in the specification by replacing the sentence of the positive and negative forms. FIG. 22 is a flowchart showing a method for replacing positive / negative words. FIG. 23 is a diagram illustrating an example of a word designating a target. FIG. 24 is a diagram illustrating an example of a document in which word combinations are replaced. FIG. 25 is a diagram illustrating an example of extraction results of word combinations including a common noun or verb.

  FIG. 1 is a schematic configuration diagram showing an embodiment of a test case creation system according to the present invention. This system includes a known computer having a control unit 10, a storage unit (storage means) 20, an output unit 30, an input unit 40, a communication unit 50, and the like.

  The control unit 10 stores various processing programs stored in the storage area of the storage unit 20, various instructions input from the input unit 40, various data corresponding to the instructions, and the like in the storage unit 20. Various processes are executed according to various processing programs stored in the storage unit 20 in accordance with various data, and the processing results are temporarily stored in the storage unit 20 and output to the output unit 30 and the like.

  As shown in FIGS. 2 and 3, the control unit 10 extracts a character string data (text specification) from electronic data of a document such as a specification and stores it in the storage unit 20 (data Extraction means) and a function for performing morphological analysis of each sentence included in the document and storing the words decomposed by the morphological analysis in the factor / viewpoint DB (first database) in association with the determined part of speech Means) and each word from the factor / viewpoint DB, and organizes the words and phrases (described later) made up of those words, and obtains words or combinations of words that satisfy a predetermined part-of-speech condition (described later) Then, replace the acquired word or combination of words with another word or combination of words within the sentence that contains them to generate a new sentence and test it as a test case Chromatography comprising generates DB function to be stored in the (second database) and (test case generation means), test case generation DB acquired the ability to output the test cases from the (output means). These functions are realized by the control unit 10 reading and executing various processing programs stored in the storage unit 20.

  The storage unit 20 includes a storage medium that stores programs, data, and the like. The storage medium is configured by a magnetic or optical recording medium or a semiconductor memory. This storage medium is fixedly provided in the storage unit 20 or is detachably mounted, and various processing programs (including a test case creation program according to the present invention) executed by the control unit 10. And a storage area for storing various databases such as a factor / viewpoint DB and a test case generation DB, and various dictionaries such as a category dictionary.

  As will be described later, the factor / point-of-view DB uses words obtained by morphological analysis to identify the type of part of speech, the corresponding factor, the point of view or clause in which the word is included, or the position (sentence line number, The test case generation DB is a database that sequentially stores test cases generated in each process to be described later.

The output unit 30 includes a CRT (Cathode Ray Tube), an LCD (Liquid Crystal Display), or the like, and displays various screens based on display data input from the control unit 10.
The input unit 40 is configured by a pointing device such as a mouse, a keyboard, and the like, and outputs an input instruction signal to the control unit 10.
The communication unit 50 is configured by a modem, a router, a bridge, and the like, and outputs data received from another computer or the like via the network to the control unit 10 and also receives data received from the control unit 10 via the network. Controls communication output to a computer or the like.

  Next, test case creation processing executed by the test case creation system will be described. As shown in FIG. 3, this test case creation process includes a text conversion step, a morpheme analysis execution step, a phrase classification step, a viewpoint generation step, and a test case generation step (extraction from the original text / word replacement / word combination replacement). Each process is sequentially performed according to a program stored in the storage unit 20 of the test case creation system.

1. Texting Step In this step, the control unit 10 first makes electronic data (for example, FIG. 4) of a document (for example, a requirement definition document, a conceptual design document, a functional specification document, or an instruction manual) in which product specifications are described. Electronic data) is read from the storage unit 20, information other than character strings is removed from the electronic data, and then a document is organized for each line. For example, when there is a punctuation mark, the line up to the punctuation mark is regarded as one sentence, and when there is no punctuation mark, the line feed code is regarded as a punctuation mark. In the case of bullets, each bulleted sentence is regarded as a single line. However, when there are a plurality of punctuation points in one itemized sentence, the sentences up to each punctuation point are considered as one sentence. As described above, the document can be made into text on a line-by-line basis. In the subsequent processing, one line is treated as one sentence, and the order of sentence description is indicated by line numbers. When the processing for organizing the document is completed, the control unit 10 performs processing for storing the organized character string data (for example, character string data as illustrated in FIG. 5) in the storage unit 20.

2. Morphological analysis execution step In this step, the control unit 10 reads character string data, performs morphological analysis of each sentence included in the document, and associates the word obtained as a result with the determined part of speech. Processing to store in the viewpoint DB is performed.

  Various well-known techniques can be used as the morphological analysis technique. In this embodiment, the IPA part-of-speech system is used as the part-of-speech system, but other part-of-speech systems may be used. For example, when a morphological analysis is performed on Example sentence 1 “Depending on the i-mode terminal on the receiving side, the URL for browsing is described in the mail text”, the sentence is converted into 17 words as shown in FIG. It is decomposed and part of speech is given to each. In the IPA part-of-speech system, there are many types of nouns such as “noun-compound word”, “noun-general”, “noun-sa-variant connection” and “noun-non-independent-general” among nouns. In this specification, the term “noun” simply includes “noun-compound word”, “noun-general”, “noun-sa-variant connection”, “noun-non-independent-general”, and the like. The “noun-compound word” indicates a combination of words in which “noun-general” and “noun-sa-variant connection” are consecutive. The same applies to other parts of speech such as “verb” and “particle”.

  Each word thus decomposed is stored in the factor / viewpoint DB in a state associated with each type of part of speech and information on the position (the line number of the sentence, the position of the word in the sentence).

3. Sentence classification step In this step, the control unit 10 classifies each sentence into phrases based on the position of the part of speech determined by the morphological analysis, and associates the information (description location information) with each word to determine factors and viewpoints. Processing to store in DB is performed.

  Specifically, first, according to the following conditions, each sentence included in the document is classified into three types of phrases, that is, a “subject phrase” including a function group when inputting to the system, and a result content of output “ It is divided into “predicate clauses” and “condition clauses” that contain content that can occur during input and output. The determination of the phrase is based on various factors.

"Subject phrase"
The part of speech is examined from the end of the sentence to the beginning of the sentence, and the viewpoint (described later) including the first detected “particle” “ga” or “ha” is set as the subject phrase. Note that the part of speech names have different classifications depending on the part of speech system to be used, such as “particulate particle-related particle”, “particle particle-case particle-general”, and so on. Determine.

Predicate clause
The phrase from the “noun-sa-variant connection” or “verb-independence” that appears first after searching from the end of the sentence to the end of the sentence is determined as a predicate clause.

`` Condition clause ''
As described above, after the subject clause and the predicate clause are determined, all character strings excluding those clauses are determined as conditional clauses. This conditional clause becomes a character string indicating a condition for the subject clause and the predicate clause. Here, there is no classification by modifiers, complements, objects, etc. shown in general Japanese syntax, and words included in one sentence are either "subject clause", "condition clause", or "predicate clause" To be included.

Next, the “condition clauses” classified as described above are further classified into “main condition clauses” and “sub-condition clauses” as shown below.
For example, as shown in FIG. 7A, when there is a “conditional clause” before the “predicate clause” and includes “noun-adverbic” or “noun-non-independent-adverbic”, the “predicate clause” A clause that does not include the following “sub-condition clause” that is located immediately before the “sub-condition clause” is defined as a “main condition clause”. The range from the word "noun-adverbable" or "noun-non-independent-adverbable" to the beginning of the sentence excluding "subject clause" and "predicate clause" is defined as "subcondition clause". .
In addition, as shown in FIG. 7B, regardless of the presence or absence of “noun-adverb” or “noun-non-independent-adverb”, a “condition” is placed before the “subject clause” and before the “predicate clause”. When there is a “clause”, the “conditional clause” positioned immediately before the “predicate clause” is set as the “main condition clause”, and the “subcondition clause” is set before the “subject clause”.
Further, as shown in FIGS. 7C and 7D, a sentence that does not include “noun-adverb” or “noun-non-independent-adverb” is available before “predicate clause” or “subject clause”. If there is a “conditional clause” in front of “,” this is the “main conditional clause”. In this case, the sentence does not include the “sub-condition clause”.

When the classification of the “condition clause” is completed, the “subject clause” and the “predicate clause” are combined to generate a “main prescribed sentence” as shown in FIGS.
When the above phrase organization is completed, the factor / point-of-view DB correlates whether each word in the sentence belongs to any of the clauses of “main prescription clause”, “main condition clause”, or “subcondition clause” To store. This information can be used as description location information indicating a phrase in which the word is used when extracting a word or registering it in a dictionary, which will be described later. In addition, by holding this description location information, it is possible to easily determine in which clause the character string to be replaced is included in the sentence at the time of replacement processing of word combinations described later. be able to.

For example, disassembling the example sentence 1 “Browsing URL is described in the mail text depending on the i-mode terminal on the receiving side” shown in FIG. “For viewing in mail text” is the main condition clause, and “depending on the receiving i-mode terminal” is the sub-condition clause.
In addition, when the example sentence 3 “If the receiving i-mode terminal does not support deco-mail with a Flash image inserted, the URL for browsing is described in the mail text.” “Is not supported for the main text,“ for viewing in the mail text ”is the main condition clause,“ the i-mode terminal on the receiving side ”is the subject of the sub-condition clause, and“ Deco-mail with a Flash image inserted ”is not supported In this case, “is a subcondition clause.

4). Viewpoint generation step In this step, the control unit 10 classifies each word stored in the factor / viewpoint DB into each factor based on the part of speech determined by the morphological analysis and the position in the sentence. The functional factors, conditional viewpoints, operational viewpoints, execution viewpoints or results viewpoints are generated by combining the same factors that are consecutive in the process, and the viewpoints are associated with each word and stored in the factor / viewpoint DB. .

  Factors are the basis of test cases such as functions and operations. In the present embodiment, five types are defined, and these are called function factors, condition factors, operation factors, execution factors, and result factors. In addition to the factor, the term “noun-adverbic” or “noun-non-independent-adverbic” is defined as a conditional phrase. For example, “when” of “when to” is a conditional phrase. In the factor classification, the factor once classified may be changed to another factor. Therefore, in the present embodiment, in the order of “functional factor”, “operation factor”, “result factor” → “execution factor” → “condition factor”. Perform the classification. Factors are determined based on various factors. An example is shown below.

(1) Function factor ・ Noun (“noun-compound word”, “noun-general”, “noun-sa-variant connection”) before “ga” or “ha” of the first “particle” from the beginning of the sentence is defined as a function factor To do.
・ If there is a part-of-speech name of “participant-combined” between nouns, all of them are regarded as functional factors.

(2) Condition factors / things that contain numerical values in unknown words are defined as condition factors.
・ Change the contents (functional factor, operation factor, execution factor, result factor) described at the head of the sentence rather than the conditional phrase to a conditional factor. However, the words (functional factors) from the beginning of the sentence to “ga” or “ha” are excluded, and the word preceding the conditional phrase is used as the conditional factor.
-If there is a “verb-independence” immediately after the conditional factor word, include the word “verb-independence”.
-The nouns, verbs, and adjectives described before the words "noun-independent-adverb", "noun-adverb", "particle-case particle-collocation" are used as conditional factors. For example, “possible adverb” includes “case”, “middle”, “after”, “when”, “when”, and the like.
・ A numerical factor (Arabic numerals / Chinese numerals) and the word immediately following are used as conditional factors. For example, “one case”, “500 bytes”, “two”, “three”, and the like correspond to this.
-Adjective words other than "adjective-self-supporting" described in the operation factor are conditional factors.

(3) Manipulation factor-The word "adjective-independence" described immediately before "noun-sa-variant connection" or "verb-independence" is used as the manipulation factor. For example, “fast”, “fast”, “slow”, and the like correspond to this.
-Words related to "adverb-participant connection" and "adverb-general", which are words representing actions and actions, are used as operation factors. For example, “slow”, “immediately”, “a little”, “pretty”, and the like correspond to this.

(4) Execution factor ・ As mentioned above, if there is a part-of-speech name of “participant-combined” between nouns, all of them are functional factors, but other “noun-sa-variant connections” are defined as execution factors. To do.
・ The word “noun-sa change connection” at the end of the sentence is used as an execution factor. For example, “edit”, “delete”, “send”, “convert”, “move”, “press”, and the like correspond to this.
・ The word “prefix-noun connection” such as “not” or “not” indicating negation is the same factor as the adjacent word.

(5) Result factor (character string that is a predicate of the main specification statement)
Search from the end of the sentence and use the word “noun-sa-variant connection” or “verb-independence” located between the first appearing particle and the end of the sentence as a result factor.
-If there is no "noun-sa-variant connection" at the end of the sentence that is not classified as another factor, the word "verb-independence" is used as the result factor. For example, “do”, “do”, “become” correspond to this.

  Words that are not classified by any of the above factors are regarded as “not subject to classification”. However, as described above, “noun-adverb” or “noun-non-independent-adverb” is a conditional phrase. By this factor classification, when there are a plurality of “ga” or “ha” in one sentence, the subject in each clause of the main specified sentence (subject and predicate), main condition clause, and subcondition clause is also specified.

  Specifically, example sentence 1 shown in FIG. 9 “Depending on the receiving i-mode terminal, a URL for browsing is described in the body of the mail.”, Example sentence 2 “However, some terminals do not have a URL for browsing. ”, Example sentence 3“ If the receiving i-mode terminal is not compatible with Deco-mail with a Flash image inserted, the URL for browsing will be written in the mail text. ” If it classifies | categorizes into, it will come to show in FIG. 10 and FIG. The factors classified in this way are stored in the factor / viewpoint DB in association with the corresponding words.

  When classification into factors is completed in this way, viewpoints are generated by combining the same consecutive factors in one sentence. If the same factor does not continue, the factor alone becomes a viewpoint. That is, from one or more functional factors, from one or more functional factors, from one or more conditional factors, from one or more operational factors, from one operational point, from one or more execution factors, from one execution point, from one or more functional factors. A result viewpoint is generated from each result factor. For example, the above-described example sentence 1 to example sentence 3 are summarized as shown in FIG. The viewpoints thus classified are stored in the factor / viewpoint DB in association with the corresponding words as shown in FIG. 12, for example. In this example, the order in which viewpoints appear in a sentence is given as a viewpoint number, and the order in which words appear in each viewpoint is given as a viewpoint number. According to this example, it is possible to easily grasp what number and which viewpoint in the sentence each word belongs to and what number the word is positioned in.

  The viewpoint thus obtained can be used for generating a test case described later, but can also be used as a test viewpoint as it is. In that case, each viewpoint may be used alone as a test viewpoint, or a priority order may be given in the order of condition viewpoint, function viewpoint, operation viewpoint, execution viewpoint, and result viewpoint, and two or more of these viewpoints may be combined. May be used as a test viewpoint. Such a test perspective can be used, for example, in exploratory testing.

5. Test case generation step As test case generation method, a sentence described in a document such as a specification is extracted and used as a test case as it is (5.1: extracted from the original sentence), and the extracted sentence Replace the word contained in the word with another word to generate a new sentence and use it as a test case (5.2: Replace word) and the combination of words (viewpoint) contained in the extracted sentence There is a method of generating a new sentence by replacing with another word combination and using it as a test case (5.3: replacement of word combination). Each of these will be described below.

5.1 Test case generation step (extracted from original text)
In this step, the control unit 10 extracts a sentence corresponding to the test case (or deletes a sentence that does not correspond to the test case) from the sentences described in the document such as the specification document, and extracts them. A process of storing a sentence as a test case in the test case generation DB is performed. When extracting as a test case, the punctuation mark is determined as the end of one sentence and stored as one test case. For a bulleted sentence, one bulleted sentence is stored as a single test case.

  In the present embodiment, a sentence that satisfies any of the following conditions (1) to (3) and satisfies the condition (4) is regarded as a sentence that does not correspond to the test case and is deleted.

(1) Sentence ending with a noun For example, if there is a sentence “priority search”, the “search” at the end of the sentence becomes “noun-sa change connection”, and this condition is met. In the case of a sentence “search preferentially”, “Yes” at the end of the sentence is “verb-independence”, so this condition is not met.
(2) Sentences in which the subject predicate relationship does not hold Sentences that do not include “particle-related particles” such as “ga” and “ha” correspond to this condition.
(3) Sentences that do not contain verbs (sentences that do not contain outcome factors)
Since the operation of the product to be tested and the operation of the user should include verbs such as “Yes”, “Done”, and “Yes”, a sentence not including this is excluded from the test target.
(4) A sentence title or item name that does not have a punctuation mark “.” At the end of the sentence satisfies this condition because the punctuation mark “.” Is not written at the end of the sentence.

  Furthermore, in this embodiment, the dependency relation between each sentence included in the factor / viewpoint DB is examined, and it is determined whether or not to be a test target based on the relation. That is, if the main sentence is the test target, the subordinate sentence is also set as the test target, and if the main sentence is not the test target, the subordinate sentence is also excluded from the test target. The dependency relationship is determined as shown in the flowchart of FIG.

First, a sentence (a combination of words) is extracted from the factor / viewpoint DB (step S101), nouns in the sentence are extracted (step S102), and the same noun word is included in consecutive sentences. Therefore, it is determined that there is continuity of meaning, and the previous sentence is “main” and the subsequent sentence is “subordinate” (steps S103 and S104).
Even if the above conditions are not met, a sentence that begins with a conjunction such as “and” “or” “and” “if” “or” “accordingly” “however” is subordinate to the immediately preceding sentence (step S105). Step S106).
Similarly, a sentence beginning with the word “noun-pronoun-general” such as “it” “this” or a sentence including these words is subordinate to the immediately preceding sentence (step S107, step S108).
For sentences starting with “from above” or “above”, nouns included in the sentence are extracted (steps S109 and S110), and sentences using these nouns are extracted from the top of the document data. Search is performed (step S111), and when the corresponding sentence exists, the sentence is mainly used as the subordinate (sentence beginning with “from above” or “from above”) (step S112).

5.2 Test case generation step (word replacement)
In this step, the control unit 10 performs a process of generating a new test case by replacing a specific word included in the sentence with another word and storing the generated test case in the test case generation DB. There are three methods for replacing words: (1) replacement of noun words, (2) replacement of positive and negative words, and (3) addition of test objects.

(1) Replacement of Noun Words In this method, a category dictionary in which words that are nouns (including nouns—compound words, nouns—generals, nouns—sa-variant connections, etc.) are registered by category is used.
In this method, after extracting a word as a noun from the factor / viewpoint DB, the control unit 10 determines whether the word exists in the category dictionary, and determines that the word exists in the category dictionary. Then, another word belonging to the same category as the word is obtained from the category dictionary, and the word is replaced with the other word in the sentence including the word, thereby generating a new sentence, Processing to store the test case in the test case generation DB is performed.

Specifically, for example, a dictionary as shown in FIG. 15 can be used as the category dictionary. In this dictionary, “JPEG image”, “FLASH image”, “GIF image”, and the like are registered in the category “image processing”.
For example, the sentence to be processed is the above-described example sentence 3 “if the receiving i-mode terminal is not compatible with deco-mail with a flash image inserted, a browsing URL is described in the mail text”. In this case, as shown in FIG. 16, all nouns such as “Flash image” are extracted (step S201), and it is determined whether or not each noun is registered in the category dictionary (step S202). Then, when registered in the category dictionary such as “Flash image”, another word (for example, “JPEG image” or “GIF image”) belonging to the same category (“image processing”) is searched. (Step S203), a new sentence is generated by replacing the “Flash image” in the sentence with another word of the same category (Step S204). As a result, a new sentence such as “If the receiving i-mode terminal is not compatible with deco-mail with a GIF image inserted, a browsing URL is described in the mail text” from Example sentence 3. Can be obtained.

  Note that the word registered in the category dictionary varies depending on the test target. For example, if the program to be tested does not clearly have the JPEG function, the word “JPEG image” is excluded from the dictionary. It is burned. The registration of the words in the category dictionary may be performed by the user, or the noun satisfying a predetermined condition may be extracted from the sentence and automatically registered in the dictionary. For example, by narrowing down with the same part-of-speech classification and searching for a character string containing the same word, registration into the category dictionary can be automatically performed. In the example of FIG. 15, by searching for a word that includes an alphabet in the compound word classification and ends with an image, the corresponding word can be extracted from the factor / viewpoint DB and automatically registered in the category dictionary. The category dictionary can be automatically generated on condition that the start word and the end word are similar or that the same word is included in the character string. The user may appropriately register words that are not described in the original document.

(2) Replacement of affirmative / negative word In this method, after the control unit 10 extracts a word representing negation from the factor / viewpoint DB based on its part of speech, the word representing the negation is In the included sentence, a new sentence is generated by replacing it with an affirmative word, and a process for storing it as a test case in the test case generation DB is performed.

  As a word representing negation, for example, as shown in FIG. 17, “adjective-case particle” “is adjective-independence” “not” or “noun-sa-variant connection” immediately before “prefix” -“Noun connection”, “non”, etc.

  When such a word representing negation is converted into a word representing affirmation and a new sentence is generated, as shown in FIG. 19, the whole sentence is negative if it is negative, and negative if positive. A new sentence is further generated by converting it into a form. Specifically, for example, a dictionary such as shown in FIG. 18 is prepared in advance, and at the time of negative / positive conversion, the word at the end of the sentence is searched in the factor / viewpoint DB, and the opposite word (negative) Is obtained from the dictionary and replaced with the word at the end of the sentence, the entire sentence is converted from the positive form to the negative form, or from the negative form to the positive form. Can do.

  For example, when the above-described example sentence 3 “if the receiving i-mode terminal does not support deco-mail with a flash image inserted, a browsing URL is written in the mail body” (see FIG. 21). In number 1) (step S301), as shown in FIG. 22, “non-corresponding” is extracted as a word representing negation in the sentence (step S302), and this is an affirmative form (in this case “corresponding”). To a new sentence (in this case, if the receiving i-mode terminal supports deco-mail with a flash image inserted, the URL for browsing is written in the mail text). ”) (Step S303), and by reversing the affirmation and negation of the generated sentence and the entire original sentence, a new sentence (in this case,“ Receiving side i-mode terminal is Fla When the deco-mail with the h image inserted is not supported, the URL for browsing is not described in the body of the mail. ”and“ 4. Receiving side i-mode terminal has inserted the Flash image ” In the case of deco-mail support, the URL for browsing is not described in the mail body. ") Is further generated (step S304), the original sentence (number 1 in FIG. 21) is" main ", and the generated sentence (figure 21 are stored as “subordinate” in the test case generation DB.

(3) Addition of test target In this method, after the control unit 10 extracts a word designating a target from the factor / viewpoint DB based on its part of speech, the target of a sentence including the word is set as another target. By replacing, a new sentence is generated and stored in the test case generation DB as a test case. For example, in the above-mentioned example sentence 1, since the description “depending on the receiving i-mode terminal” indicates the existence of “other terminals”, a test case for testing another terminal is derived from this example sentence 1. Can be made.

  Specifically, the word designating the target can include “particle-case particle-collocation”, and the target of the sentence including this word can be replaced with another target as shown in FIG. . In other words, when “by”, “in”, “about”, “introduction”, “related”, etc., “particle” -case particle-collocation ”is included, and“ particle ”-coordinate particle” is also immediately followed by this A new sentence is generated by deleting “mo” and adding the word “other” at the beginning of the sentence. On the other hand, when there is “ha” in “particle-clause particle” immediately after “particle-case particle-collocation”, this “ha” is deleted and the word “other” is added to the beginning of the sentence. A new sentence is generated by converting the whole sentence into a positive form if it is negative, and a negative form if it is positive. For example, from the above-described example sentence 1, a sentence such as “browsing URL is not described in the mail text by another receiving side i-mode terminal” can be newly generated as a test case.

5.3 Test case generation step (replacement of word combinations)
In this step, the control unit 10 takes out word combinations classified as the same type of viewpoint from the factor / viewpoint DB, and selects combinations of words satisfying a predetermined condition by paying attention to each part of speech composition. After the extraction, the sentence combinations including the word combinations are replaced with each other to generate new sentences and store them in the test case generation DB as test cases. In this processing, the viewpoint as a single factor in which the same factor is not continuous is not subject to replacement, but the viewpoint in which the same consecutive factors are combined is targeted.

  Specifically, first, one of the viewpoints generated in the viewpoint generation step described above (for example, a functional viewpoint) is selected, and a combination of words corresponding to the viewpoint is extracted from the factor / viewpoint DB, Then, a combination of words including common nouns or verbs is extracted and stored in the storage unit 20. The range to be extracted may be an entire document such as a specification or a specific chapter section in the document. As a combination of words including a noun or a verb, for example, from a functional viewpoint, a combination of words starting with a noun (noun 1), including a part of speech other than the noun or verb and ending with a noun (noun 2) From the viewpoint of results, there are combinations of words that start with a noun (noun 3), include a part of speech other than a noun or verb, and end with a verb (verb 1). In the latter case, a combination of words in which either one of noun 3 and verb 1 is the same is a combination of words containing a common noun or verb. Extract as a combination. Also, here, for example, verbs with different usage forms such as the words “Yes” and “Shi” in the word combinations “Reply quoted” and “Transfer not quoted” are considered to be the same verb. To be extracted. In each viewpoint in which the same consecutive factors are combined, the part of speech of the start word and the part of speech of the last word are processed in the same manner as described above.

  For example, from a document as shown in FIG. 24, a combination (viewpoint) of words including common nouns or verbs as shown in FIG. 25 is extracted. In this example, “Role List” “Role Succession” “Role Specification” “Role Details” “Role Settings” “Role Settings Screen” etc. are extracted as combinations of words that begin with the noun “role” Has been. The combination of these words exists in another sentence (for example, “Role List” exists in the sentence on the first line, “Succession of Role” exists in the sentence on the second line, etc.) Not only in another clause in one sentence (for example, in the main clause of a sentence with one of “Role List” or “Success of Role”, and in the main condition clause of the same sentence in the other sentence) In the same clause in one sentence (for example, in the main condition clause of a sentence having both “role list” and “inherit role”).

  When a combination of words including a common noun or verb is extracted in this way, a new sentence is generated by substituting the combination of the words with each other between sentences including the combination of the words. For example, in a document, there is a sentence "... list of roles ... delete update date ...", and it contains words that contain nouns or verbs in common with "list of roles". When “Role succession” is extracted as a combination, replace “Role list” with “Role succession” in the sentence, and “... A new sentence "Delete ..." can be generated. Such processing is performed for all clauses (main specification sentence, main condition clause and subcondition clause) and all viewpoints (functional viewpoint, condition viewpoint, operation viewpoint, execution viewpoint and result viewpoint). Further, a new sentence can be further generated from the new sentence generated in this manner by replacing the above-mentioned positive form / negative form word. Similarly, the generated sentence is stored in the test case generation DB as a test case.

6). Output Step In this step, the control unit 10 extracts the test cases (sentences) generated in the test case generation step from the test case generation DB, lists the test cases, and adds “test” to each of the test cases. In addition to adding items such as “Execution Date” and “Test Executor Name”, insert a sentence such as “Check this behavior” at the end of the test case sentence and use it as a test case. Output in a way that can be understood by the user. The output form may be output as an image on a display device, or may be printed on a paper medium and output.

  As described above, according to the present embodiment, the control unit 10 performs morphological analysis using character string data of a document such as a specification, and the part of speech set in advance based on the result of the morphological analysis. Obtain a word or combination of words that satisfies the condition, and generate a new sentence by replacing the obtained word or combination of words with another word or combination of words within the sentence that contains them. Is output from the output unit 30 as a test case, so that the time and labor required to create the test case can be greatly reduced, and it is possible to prevent bias in the selection of the test case, As a result, a more comprehensive and efficient test can be realized.

DESCRIPTION OF SYMBOLS 10 Control part 20 Memory | storage part 30 Output part 40 Input part 50 Communication part

Claims (16)

A system for creating test cases from documents that describe product specifications,
Storage means for storing character string data of the document;
The data of the character string is read from the storage means, morphological analysis of each sentence included in the document is performed, and the character or character string of the minimum unit obtained by the morphological analysis is used as a word, and these words are converted into the morpheme. Morphological analysis execution means for storing in the first database in association with the part of speech determined by the analysis;
A word or combination of words satisfying a preset part-of-speech condition is acquired from the first database, and the acquired word or combination of words is replaced with another word or combination of words in a sentence including them. A test case generation means for generating a new sentence and storing it as a test case in the second database;
Output means for obtaining and outputting a test case from the second database , and further,
Based on the position relationship of parts of speech determined by the morphological analysis, each sentence included in the document is divided into a main prescription sentence in which a subject clause and a predicate clause are combined, and one or a plurality of conditional clauses excluding the main prescription sentence; Clause classification means for storing the information in the first database in association with each word constituting the clauses;
The viewpoints related to the function, condition, operation, execution and result of the test are the functional viewpoint, the conditional viewpoint, the operation viewpoint, the execution viewpoint and the result viewpoint, respectively. When a factor, an execution factor, and a result factor are used, each word stored in the first database is classified into one of the factors based on the part of speech determined by the morphological analysis and the position in the sentence. Factor classification means for associating the classification result with each word and storing it in the first database;
If the same factor continues in a sentence, the combined factors are generated, and if not, a single factor is generated as a functional viewpoint, a conditional viewpoint, an operational viewpoint, an execution viewpoint, or a results viewpoint, Viewpoint generating means for associating those viewpoints with each word and storing them in the first database,
The test case generation means extracts a combination of words classified as the same kind of viewpoint from the first database, extracts a combination of words including a common noun or verb, and then extracts A test case creation system characterized in that a new sentence is generated by substituting a combination of words among sentences including a combination, and stored as a test case in a second database . A system for creating test cases from documents that describe product specifications,
Storage means for storing character string data of the document;
The data of the character string is read from the storage means, morphological analysis of each sentence included in the document is performed, and the character or character string of the minimum unit obtained by the morphological analysis is used as a word, and these words are converted into the morpheme. Morphological analysis execution means for storing in the first database in association with the part of speech determined by the analysis;
A word or combination of words satisfying a preset part-of-speech condition is acquired from the first database, and the acquired word or combination of words is replaced with another word or combination of words in a sentence including them. A test case generation means for generating a new sentence and storing it as a test case in the second database;
Output means for obtaining and outputting a test case from the second database , and further,
It has a category dictionary in which words that are nouns are registered by category,
When the test case generation unit extracts a word as a noun from the first database, determines whether the word exists in the category dictionary, and determines that the word exists in the category dictionary Then, another word belonging to the same category as the word is obtained from the category dictionary, and the word is replaced with the other word in the sentence including the word, thereby generating a new sentence, Is stored in the second database as a test case. A dependency relation between sentences included in the document is determined, and as a result of the determination, if the main sentence satisfies a predetermined condition used as a test case, the dependent sentence is also stored in the second database as a test case. 3. The test case creation system according to claim 1, wherein, if the main sentence does not satisfy the predetermined condition, the subordinate sentence is also excluded from the test case. The nouns in the sentence are extracted, and when the same noun word is included in consecutive sentences, the previous sentence is determined as a main sentence and the subsequent sentence is determined as a subordinate sentence, respectively. The test case creation system according to claim 3 . 4. The test case creation system according to claim 3 , wherein a sentence starting with a conjunction is determined as a sentence subordinate to the immediately preceding sentence. 4. The test case creation system according to claim 3 , wherein a sentence beginning with the word “noun-pronoun-general” or a sentence including the word is determined as a sentence subordinate to the immediately preceding sentence. Extract nouns contained in sentences starting with “From”, “From” or their synonyms, and search for the sentences in which those nouns are used from the beginning of the document data. 4. The test according to claim 3 , wherein the sentence is determined as a subordinate sentence, with the sentence as a main sentence, and sentences beginning with “from” or “above” or synonyms thereof, respectively. Case creation system. The test case generation unit further extracts a word representing negation from the first database based on the part of speech, and then replaces the word representing negation with a word representing affirmation in a sentence including the word. it allows test case creation system according to claim 1 or 2 to generate a new sentence, characterized by storing in said second database it as a test case. The test case generation means, for a sentence including the word indicating the negative and a new sentence generated from the sentence, if the sentence is a negative sentence, an affirmative sentence 9. A test case creation system according to claim 8 , wherein new sentences are further generated by converting them into negative forms when they are, and are stored in the second database as test cases. . The test case generating means further from the first database, after extracting based word specifying a target part of speech, in sentence containing the word, before the subject to which the word is to specify "other" The test case creation system according to claim 1 or 2 , wherein a new sentence is generated by adding the word "" and stored in the second database as a test case. It extracts data string from the electronic data of the document, test case creation system according to claim 1 or 2, further comprising a data extraction means for storing in the storage means. The test case creation system according to claim 1, further comprising a test viewpoint output unit that outputs the generated viewpoint or a combination thereof as a test viewpoint . Executed by a computer comprising storage means for storing character string data of a document in which product specifications are described, and control means for performing control to generate a test case from the character string data stored in the storage means A test case creation method,
The control unit reads the data of the character string from the storage unit, performs morphological analysis of each sentence included in the document, and uses the character or character string of the minimum unit obtained by the morphological analysis as a word. Storing the word in the first database in association with the part of speech determined by the morphological analysis;
The control means acquires a word or a combination of words satisfying a preset part-of-speech condition from the first database, and the acquired word or word combination is changed to another word or Generating a new sentence by replacing it with a combination of words and storing it in a second database as a test case;
The control means includes obtaining a test case from the second database and outputting the test case ;
Based on the positional relationship of the part of speech determined by the morphological analysis, the control means includes, for each sentence included in the document, a main prescription sentence that combines a subject clause and a predicate clause, and 1 or Dividing the information into a plurality of conditional clauses, and storing the information in the first database in association with each word constituting the clauses;
The viewpoints related to the function, condition, operation, execution and result of the test are the functional viewpoint, the conditional viewpoint, the operation viewpoint, the execution viewpoint and the result viewpoint, respectively. In the case of a factor, an execution factor, and a result factor, the control means determines that each of the words stored in the first database is the part of speech determined by the morphological analysis and the position in the sentence. And classifying the classification result with each word and storing it in the first database,
When the same factor is continued in one sentence, the control means is a combination of the factors, and when not, a single factor is a function point of view, a condition point of view, an operation point of view, an execution point of view or a result point of view. And store them in the first database in association with their respective words,
The control means takes out a combination of words classified as the same type of viewpoint from the first database, extracts a combination of words including a common noun or verb from the combination, and then the combination of the words is A test case creation method characterized in that a new sentence is generated by substituting a combination of words among contained sentences and stored as a test case in a second database . Storage means for storing character string data of a document in which product specifications are described, control means for performing control to generate a test case from the character string data stored in the storage means, and a noun word A test case creation method executed by a computer including a category dictionary registered by category ,
The control unit reads the data of the character string from the storage unit, performs morphological analysis of each sentence included in the document, and uses the character or character string of the minimum unit obtained by the morphological analysis as a word. Storing the word in the first database in association with the part of speech determined by the morphological analysis;
The control means acquires a word or a combination of words satisfying a preset part-of-speech condition from the first database, and the acquired word or word combination is changed to another word or Generating a new sentence by replacing it with a combination of words and storing it in a second database as a test case;
The control means includes obtaining a test case from the second database and outputting the test case ;
After the control means has extracted a noun word from the first database, it is determined whether the word is present in the category dictionary, and when it is determined that the word is present in the category dictionary, Get another word from the category dictionary that belongs to the same category as the word, replace the word with the other word in the sentence that contains it, generate a new sentence, and test it A test case creation method characterized in that the test case is stored in the second database as a case. Executed by a computer comprising storage means for storing character string data of a document in which product specifications are described, and control means for performing control to generate a test case from the character string data stored in the storage means A program
The control unit reads the data of the character string from the storage unit, performs morphological analysis of each sentence included in the document, and uses the character or character string of the minimum unit obtained by the morphological analysis as a word. Storing the word in the first database in association with the part of speech determined by the morphological analysis;
The control means acquires a word or a combination of words satisfying a preset part-of-speech condition from the first database, and the acquired word or word combination is changed to another word or Generating a new sentence by replacing it with a combination of words and storing it in a second database as a test case;
The control means causes the computer to execute a step of acquiring and outputting a test case from the second database ; and
Based on the positional relationship of the part of speech determined by the morphological analysis, the control means includes, for each sentence included in the document, a main prescription sentence that combines a subject clause and a predicate clause, and 1 or Dividing the information into a plurality of conditional clauses and storing the information in the first database in association with each word constituting the clauses;
The viewpoints related to the function, condition, operation, execution and result of the test are the functional viewpoint, the conditional viewpoint, the operation viewpoint, the execution viewpoint and the result viewpoint, respectively. In the case of a factor, an execution factor, and a result factor, the control means determines that each of the words stored in the first database is the part of speech determined by the morphological analysis and the position in the sentence. And storing the classification result in the first database in association with each word;
When the same factor is continued in one sentence, the control means is a combination of the factors, and when not, a single factor is a function point of view, a condition point of view, an operation point of view, an execution point of view or a result point of view. And associating them with the respective words and storing them in the first database;
The control means takes out a combination of words classified as the same type of viewpoint from the first database, extracts a combination of words including a common noun or verb from the combination, and then the combination of the words is A program that causes a computer to execute a step of generating a new sentence by replacing a combination of words among included sentences and storing the new sentence as a test case in a second database . Storage means for storing character string data of a document in which product specifications are described, control means for performing control to generate a test case from the character string data stored in the storage means, and a noun word A program executed by a computer comprising a category dictionary registered by category ,
The control unit reads the data of the character string from the storage unit, performs morphological analysis of each sentence included in the document, and uses the character or character string of the minimum unit obtained by the morphological analysis as a word. Storing the word in the first database in association with the part of speech determined by the morphological analysis;
The control means acquires a word or a combination of words satisfying a preset part-of-speech condition from the first database, and the acquired word or word combination is changed to another word or Generating a new sentence by replacing it with a combination of words and storing it in a second database as a test case;
The control means causes the computer to execute a step of acquiring and outputting a test case from the second database ; and
After the control means has extracted a noun word from the first database, it is determined whether the word is present in the category dictionary, and when it is determined that the word is present in the category dictionary, Get another word from the category dictionary that belongs to the same category as the word, replace the word with the other word in the sentence that contains it, generate a new sentence, and test it A program for causing a computer to execute the step of storing in the second database as a case .

Images