QA AI Project:VS Code, Cursor, & AI for API/UI Testing | Building FoXYiZ LCNC Framework скачать в хорошем качестве

QA AI Project:VS Code, Cursor, & AI for API/UI Testing | Building FoXYiZ LCNC Framework

Visit our new itelearn.com; built and powered through AI for members. This session focuses on taking test automation to the next world using AI, building on the orientation conducted over the previous weekend. The primary focus is project work, including the development of FoXYiZ, a low code no code automation framework. We are building the FoXYiZ app as a cloud-based solution, eliminating the need for a desktop version. Participants are encouraged to gain expertise in this project and carry it forward in their professional profiles. Key Progress Review & Demos: • API Testing with Cloud AI: The session includes a review of progress made using VS Code integrated with Cloud AI. A participant demonstrated generating API tests, including fixing initial errors. ◦ The test results were used to create detailed HTML reports and spreadsheet reports. ◦ Reports included critical details like response codes (e.g., expecting 200 but receiving 404) and time taken, which is crucial for stakeholders. ◦ Building these tests without AI assistance would typically take several days or a week. • UI Testing with Playwright: UI testing progress was demonstrated using Playwright (not Playwright MCP) to create login and registration tests. This testing also generated HTML reports and Allure reports. Technical Deep Dives & Discussion: • Context Engineering in Cursor: A crucial topic covered was Context Engineering in the Cursor IDE. This allows developers to add specific project folders (like 'pi scripts' or 'demo webshop tentis') to a dedicated chat context, preventing the AI from getting confused across multiple projects. • Tool Integration and Collaboration: We discussed how to integrate different IDEs (like VS Code, PyCharm, and Cursor). The suggested best practice for remote teams is using a central repository like GitHub or BitBucket. • AI Models and Costs: The conversation touched on the high cost of AI processing due to reliance on mega GPU processing hardware, which leads to limitations like running out of free tokens in platforms like Cursor. Different models were noted, including GPT-5 Mini, GPT-4o, Gemini 2.5, and Cheetah. The measurement of consumption is based on the number of tokens, not the number of prompts. • Vibe Coding Philosophy: The goal is to maximize "vibe coding" (having AI write the code) so engineers can smartly distribute their time to test, refine, and improve the AI-generated code. The philosophy discussed encourages engineers to move away from manually writing code, as they already possess the necessary historical knowledge of how code works. The aim is to use AI to accomplish almost the entire project task. AI is used to assist with documentation, script/test creation, data creation, execution, result analysis, self-healing, GitHub integration, and the entire CI/CD pipeline and deployments. (Note: Earlier sessions covering Python fundamentals and the AI concepts, including "Vibe coding with Python," are available on the ITeLearn website under the AI tab, specifically sessions from September 20th and 21st). Next Steps & Resources: • We plan to integrate GitHub more deeply (including making the current public repository private and adding contributors). Integrating GitHub is the next step toward implementing CI/CD pipelines using tools like Jenkins. • The code and documentation will be shared on the ITeLearn website within the "QA to AI" study stream. Earlier Python fundamentals and Vibe coding sessions are available under the AI tab on the website (specifically the sessions from September 20th and 21st). • We will continue exploring tools like Replit and Augment Code if token limitations persist with Cursor