20.500.12556/RUL-172524 Uporaba s poizvedovanjem obogatenega generiranja za avtomatsko generiranje testov Using retrieval-augmented generation for automated test generation Pisanje testov je časovno zahteven del razvoja programske opreme in razvijalci vsak dan porabijo veliko časa za njihovo pripravo. Ob hitrem razvoju na področju velikih jezikovnih modelov smo hoteli preveriti uspešnost tehnologije s poizvedovanjem obogatenega generiranja (RAG) za avtomatsko generiranje testov v jeziku Python. Veliki jezikovni modeli so uspešni pri generiranju kode, a kakovostno testiranje je ena izmed njihovih večjih pomanjkljivosti, pogosto namreč generirajo teste brez razumevanja konteksta. RAG je tehnika, ki zaobide pomanjkanje domenskega znanja tako, da na podlagi analizirane funkcije v bazi znanja najde podobne teste, s katerimi lahko pravilno generira nove. Z najdenimi primeri jezikovni model ustvari teste brez halucinacij in drugih omejitev. Raziskali in implementirali smo prototip sistema in ga preizkusili na realnem projektu z objektivnimi in subjektivnimi metrikami. Generirani testi so dosegli višjo pokritost kode, medtem ko so bili pri odkrivanju dejanskih napak manj uspešni od ročno napisanih testov. Programerji pa so s subjektivnimi ocenami dodatno potrdili prednost ročnih testov. Writing tests is a time-consuming part of software development, and developers spend considerable time preparing them daily. With rapid advances in large language models, we investigated the effectiveness of Retrieval-Augmented Generation (RAG) technology for automated Python test generation. Large language models are successful at generating code, but quality testing is one of their major weaknesses; they often generate tests without understanding context. RAG is a technique that bypasses the lack of domain knowledge by identifying similar tests in a knowledge base based on the analyzed function, which it uses to generate new ones correctly. With the retrieved examples, the language model creates tests without hallucinations and other limitations. We researched and implemented a prototype system and tested it on a real project using objective and subjective metrics. Generated tests achieved higher code coverage, while being less successful at detecting actual bugs compared to manually written tests. Developers further confirmed the superiority of manual tests through subjective assessments. avtomatsko generiranje testov veliki jezikovni modeli testiranje programske opreme RAG automated test generation RAG large language models software testing. true false false Slovenski jezik Angleški jezik Diplomsko delo/naloga 2025-09-08 10:40:01 2025-09-08 10:40:09 2025-10-15 03:39:57 0000-00-00 00:00:00 2025 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 38184 248932611 Starasinic_Zan_-_Uporaba_s_poizvedovanjem_obogatenega_generiranja_za_avtomatsko_generiranje_test.pdf Starasinic_Zan_-_Uporaba_s_poizvedovanjem_obogatenega_generiranja_za_avtomatsko_generiranje_test.pdf 1 C5BC8A8817A9C43177090CF3CF68FB88 5eb26baf1538d858cd96602dc8a331b924c7eff1f0d7d81183b83d2f22fe049c 477ba836-8c8f-11f0-9328-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=216354 Fakulteta za računalništvo in informatiko 0 0 0