Semantic Test Suite Reduction via Vector Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Enterprise application test suites become inefficient and ineffective due to their large size, poor management, and redundancy, leading to unnecessary resource expenditure and lower software quality as teams unknowingly test the same functionalities multiple times.
Innovation Solution
A method involving a test management computing device that uses semantic vectorization to generate a vector model, implements cluster optimization to identify maximally separate and compact clusters, and determines a subset of test scripts for reduction, thereby optimizing test suites by removing redundant scripts while maintaining code coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If test suites are expanded to cover more functionalities, then code coverage is improved, but redundancy increases and testing efficiency deteriorates
Solution Approach 1:
The patent extracts and removes redundant test scripts from the test suite by comparing semantic similarities. The system identifies test scripts that test the same functionality and removes duplicates, thereby maintaining code coverage while eliminating redundancy that harms testing efficiency.
Solution Approach 2:
The patent transforms test scripts into semantic vectors and uses clustering algorithms to group similar tests. By changing the representation parameter from raw text to semantic vectors, the system can efficiently identify and remove redundant tests while preserving necessary coverage.
2Ease of manufacture
If test suites are manually managed by individual testers, then test script creation is flexible, but homogeneity and understanding of the entire suite deteriorate
Solution Approach 1:
The patent introduces an intermediary system that automatically analyzes test scripts using semantic vectorization and clustering. This intermediary process mediates between individually created test scripts and the overall suite management, providing holistic understanding and identifying redundancies without restricting individual creativity.
Solution Approach 2:
The system provides feedback about test suite redundancy and quality by analyzing semantic similarities. This feedback mechanism helps maintain homogeneity and understanding of the entire suite while preserving the flexibility of individual test script creation.
3Productivity
If semantic vectorization and cluster optimization are applied to identify redundant test scripts, then testing efficiency is improved, but computational resources are consumed
Solution Approach 1:
The patent applies semantic vectorization and clustering only to identify redundant test scripts rather than processing all test scripts equally. By focusing computational resources on detecting redundancies rather than executing all tests, the system improves testing efficiency while managing computational resource consumption.
Data Source
AI summary
Methods, non-transitory computer readable media, test management computing devices that obtain test scripts associated with a test suite for testing an application for the test scripts. A vector model is generated based on a semantic vectorization of the obtained test scripts. A cluster optimization is implemented on the vector model to identify a plurality of maximally separate and compact clusters. A subset of the test scripts that are candidates for facilitating reduction of the test suite is determined, based on the identified clusters, and an indication of each test script of the subset of the test scripts is output. With this technology, a semantic analysis of test scripts of a test suite is implemented to reduce the size of the test suite while advantageously maintaining the coverage with respect to an associated enterprise application as well as ensuring a low level of redundancy present in the test suite.