How thebuggybug/vscode_ext works
Project type and purpose
• This is a Visual Studio Code extension (“Deepseek”) written in TypeScript.
• Purpose: Provide an in-editor chatbot UI that communicates with the Ollama local AI model (deepseek-coder-v2).
File structure
• Top-level files:
• package.json, tsconfig.json, eslint.config.mjs, README.md, CHANGELOG.md, License.md
• vsce and VS Code configuration files (.gitignore, .vscodeignore, vsc-extension-quickstart.md)
• src/ directory (TypeScript source):
• extension.ts – Extension activation, command registration, lifecycle management
• panel.ts – Webview panel manager, message handling between extension and UI
• drafts.js – (possibly helper or experimental code)
• test/extension.test.ts – Unit tests for extension activation and message handling
• src/webview/ – Static assets for the embedded chatbot UI
• index.html, script.js, style.css
README for additional insights
• The user invokes the “Start Deepseek” command via the Command Palette.
• The extension spins up a Webview-based chatbot UI in a VS Code panel.
• Communications flow over VS Code’s postMessage API from webview → extension host.
• The extension invokes the Ollama CLI (child_process.spawn or similar) to pass user messages to the local deepseek-coder-v2 model, receives responses, and relays them back to the webview.