Test case generation from requirements
A method, a device, and a non-transitory storage medium are described in which a test case generation service is provided. The service may include receiving and validating requirement text associated with a prospective test case and/or test script. The service may invoke a remedial procedure when the requirement text is not validated. The service may include selecting multiple types of classification models based on a model profile. The service may include aggregating probability values associated with classifications for sentences, and approving such classification that satisfy a threshold value. The service may further include generating test cases and/or test scripts. The service may generate a schedule for test cases and/or test scripts based on one or multiple criteria values generated by one or multiple predictive models.
1 . A method comprising:
analyzing, by a device, a requirement text relating to at least one of a test case or a test script;
selecting, by the device based on the analyzing and in response to determining that the requirement text does not include an error, a model profile, wherein the model profile indicates a resource usage criterion to satisfy by each classification model before permitted to classify each sentence of the requirement text;
determining, by the device, a classification for each sentence of the requirement text based on classifying each sentence using multiple types of classification models that satisfied the model profile, wherein the multiple types of classification models include a tree-based classification model, a vector-based classification model, and a transformer-based classification model;
aggregating, by the device, each probability value relating to each classification associated with each sentence provided by each of the multiple types of classification models;
selecting, by the device based on the aggregating, a first classification for each sentence of the requirement text; and
generating, by the device based on the selecting, the at least one of the test case or the test script.
2 . The method of claim 1 , wherein the error includes two or more of a typographical error, an incorrect labeling of one or more sentences of the requirement text, or an insufficient number of sentences.
3 . The method of claim 1 , wherein the aggregating comprises:
averaging, by the device, each of the probability values associated with each classification associated with each sentence.
4 . The method of claim 1 , further comprising:
comparing, by the device, an aggregated probability value to a threshold probability value; and
determining, by the device, whether the aggregated probability value satisfies the threshold probability value.
5 . The method of claim 1 , further comprising:
determining, by the device, that the requirement text includes an error; and
invoking, by the device, a remedial procedure.
6 . The method of claim 1 , wherein the model profile further indicates a utility criteria.
7 . The method of claim 1 , further comprising:
comparing, by the device, the model profile to a corresponding resource usage criteria of the multiple types of classification models.
8 . The method of claim 1 , further comprising:
generating, by the device, a schedule to execute the at least one of the test case or the test script based on one or multiple criteria values generated by one or multiple predictive models.
9 . A device comprising:
a processor configured to:
analyze a requirement text relating to at least one of a test case or a test script;
select, based on the analysis and in response to determining that the requirement text does not include an error, a model profile, wherein the model profile indicates a resource usage criterion to satisfy by each classification model before permitted to classify each sentence of the requirement text;
determine, a classification for each sentence of the requirement text based on classifying each sentence using multiple types of classification models that satisfied the model profile, wherein the multiple types of classification models include a tree-based classification model, a vector-based classification model, and a transformer-based classification model;
aggregate each probability value relating to each classification associated with each sentence provided by each of the multiple types of classification models;
select, based on the aggregation, a first classification for each sentence of the requirement text; and
generate, based on the selection, the at least one of the test case or the test script.
10 . The device of claim 9 , wherein the error includes two or more of a typographical error, an incorrect labeling of one or more sentences of the requirement text, or an insufficient number of sentences.
11 . The device of claim 9 , wherein to aggregate, the processor is further configured to:
average each of the probability values associated with each classification associated with each sentence.
12 . The device of claim 9 , wherein the processor is further configured to:
compare an aggregated probability value to a threshold probability value; and
determine whether the aggregated probability value satisfies the threshold probability value.
13 . The device of claim 9 , wherein the processor is further configured to:
determine that the requirement text includes an error; and
invoke a remedial procedure.
14 . The device of claim 9 , wherein the model profile further indicates a utility criteria.
15 . The device of claim 9 , wherein the processor is further configured to:
compare the model profile to a corresponding resource usage criteria of the multiple types of classification models.
16 . The device of claim 9 , wherein the processor is further configured to:
generate a schedule to execute the at least one of the test case or the test script based on one or multiple criteria values generated by one or multiple predictive models.
17 . A non-transitory computer-readable storage medium storing instructions executable by a processor of a device, wherein the instructions are configured to:
analyze a requirement text relating to at least one of a test case or a test script;
select, based on the analysis and in response to determining that the requirement text does not include an error, a model profile, wherein the model profile indicates a resource usage criterion to satisfy by each classification model before permitted to classify each sentence of the requirement text;
determine, a classification for each sentence of the requirement text based on classifying each sentence using multiple types of classification models that satisfied the model profile, wherein the multiple types of classification models include a tree-based classification model, a vector-based classification model, and a transformer-based classification model;
aggregate each probability value relating to each classification associated with each sentence provided by each of the multiple types of classification models;
select, based on the aggregation, a first classification for each sentence of the requirement text; and
generate, based on the selection, the at least one of the test case or the test script.
18 . The non-transitory computer-readable storage medium of claim 17 , wherein the error includes two or more of a typographical error, an incorrect labeling of one or more sentences of the requirement text, or an insufficient number of sentences.
19 . The non-transitory computer-readable storage medium of claim 17 , wherein the model profile further indicates a utility criteria.
20 . The non-transitory computer-readable storage medium of claim 17 , wherein the instructions are further configured to:
compare an aggregated probability value to a threshold probability value; and
determine whether the aggregated probability value satisfies the threshold probability value.