Computing Complex for Automated Control of the Process of Developing Software for Recognition of Structured Documents
摘要
The development of software for solving applied problems involves the implementation of many routine computational experiments. The approach of continuous software integration, widely used in practice, involves frequent centralized assembly of the program and bench testing. This generates a large amount of test results, which should be promptly available to software developers for error analysis and version comparison. To automate these procedures, the architecture of the automatic quality control system for the structured document recognition program was developed, including the collection, storage and display of bench test results. Software test results are recorded in a database for further analysis. Standing software is performed on virtual servers running various operating systems. For stability, the web server and database are physically separated from the build server. Web technologies are used both to automatically upload test results to a database and to serve user requests.