3 results about "Standard result" patented technology

This invention provides a method, apparatus, and device for automatic conversion of API result codes. The method includes: implementing the CommandLineRunner interface to perform a data refresh upon project startup, converting the result codes into data stored in memory; defining static variables to store the API return result description key value and the API return result encoding key value in memory; periodically reading or manually refreshing the memory data to update the pre-stored result code data in memory; and before controlling the API to finally return the result, determining whether error code conversion is enabled. If not enabled, no conversion is performed; if enabled, regular expression matching is performed on the result codes to convert them into standard result codes and result descriptions for return. In this way, automatic conversion of error codes and descriptions is achieved by customizing whether error code conversion is enabled, configuring dictionary table data, and enabling external error code conversion.

The invention belongs to the technical field of software testing, and discloses an airport operation index comparison testing method and system based on a large language model. The method comprises the following steps: constructing a standard case set in a standard case set construction module on the basis of collected operation index data; preliminary checking of the automatic inquiry test condition is completed; inquiry result acquisition is completed in a result acquisition module; inputting the inquiry result in the result acquisition module and the case set standard result in the standard case set construction module into a result comparison script for comparison; and a visual test report is automatically generated. According to the method, the time consumed by batch scene testing is shortened from the manual hour level to the minute level, automatic execution of daily average ten thousand level inquiry cases is supported, the test comprehensiveness is improved, numerical and text type indexes are covered, numerical error calculation and semantic similarity analysis are combined, the deviation between LLM output and a standard result is quantified, and accurate comparison is achieved.

The embodiment of the invention discloses a test method and test equipment of an application interface and a computer readable storage medium. The method comprises the steps of obtaining a target picture and an initial prompt word set of a target application interface, analyzing the target picture by a target multi-modal large model in response to the initial prompt word set to obtain a test result set, screening out a target visual question and answer task according to the test result set and a preset standard result set, and displaying the target visual question and answer task. And determining a composite prompt word set corresponding to the target visual question and answer task, and transmitting the composite prompt word set to the target multi-modal large model, so that the target multi-modal large model responds to the composite prompt word set to test the target application interface. The target multi-modal large model provided by the embodiment supports a question-asking mode of cue words to output the answer result of the visual question-answering task, even if the version of the target application interface is iterated, the visual question-answering task of the UI test is not easy to change, which means that the cue words corresponding to the visual question-answering task are not changed, and the user experience is improved. The embodiment of the invention has relatively high reusability.