职位描述
I’m building an AI-powered helper inside Visual Studio Code that can turn a plain-language request into a fully-formed Python project. When I ask for “a Flask API with two endpoints” or “a Pygame tic-tac-toe game,” the extension should create every required folder and file, fill them with syntactically correct Python, and drop the new structure straight into the open workspace.
Your task is to design and implement this VS Code extension. The assistant’s first and only focus is Python, and its primary skill is writing and completing code—not just snippets, but whole modules and project scaffolds generated from a single instruction. The flow I’m after is:
• I issue a request (via command palette, sidebar, or another intuitive entry point).
• The extension calls an LLM such as GPT-4, streams back the response, and writes the output to disk in real time.
• Generated code is properly formatted, includes needed imports, and runs without modification.
Deliverables
• Source code for the VS Code extension plus a packaged .vsix file.
• README explaining install, configuration of the LLM API key, and usage examples.
• A short screen-recorded demo (GIF or MP4) showing a prompt and the resulting project scaffold.
Acceptance criteria
1. A test prompt “create a basic Flask ‘Hello, World!’ app” produces a runnable project that starts with `python [login to view URL]`.
2. A prompt with nested structure (“build a package named utils with a [login to view URL] module”) creates correct folders and `[login to view URL]` files.
3. No existing workspace files are overwritten without confirmation.
Experience with TypeScript, VS Code extension APIs, and OpenAI (or equivalent) SDKs is essential. If you can hand me a clean, documented extension that meets the above checkpoints, let’s get started.
---
來源:Freelancer.com | 出價次數:11 | 剩餘時間:7 日 | 專案類型:固定預算