DevMind

AI-powered developer tools providing code search, bug detection, code explanation, and unit test generation — all inside VS Code.

Built with a FastAPI backend and a VS Code extension, using mock data so the entire system is runnable immediately without trained models or API keys.

---

Quick Start

1. Run the Backend

cd backend

# Create and activate a virtual environment
python -m venv venv
# Windows:
.\venv\Scripts\activate
# macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Start the server
uvicorn app.main:app --reload

The API will be available at http://localhost:8000. Open http://localhost:8000/docs for interactive Swagger documentation.

2. Run the VS Code Extension

cd extension

# Install dependencies
npm install

# Compile TypeScript
npm run compile

Then:

1. Open the extension/ folder in VS Code
2. Press F5 to launch the Extension Development Host
3. Open any .py or .js file to activate the extension

3. Try It Out

Feature	How to trigger	What happens
🧠 Code Explanation	Hover over any word in a .py/.js file	Tooltip with AI explanation
🐛 Bug Detection	Save any .py/.js file	Squiggly underlines + Problems panel
🧪 Test Generation	Ctrl+Shift+P → "Code Assistant: Generate Tests"	New tab with generated tests

---

Project Structure

DevMind/
├── backend/                      ← FastAPI backend (Python)
│   ├── app/
│   │   ├── main.py               ← App entry point + CORS + router wiring
│   │   ├── api/                  ← HTTP endpoint routers
│   │   │   ├── search.py         ← POST /search
│   │   │   ├── bugs.py           ← POST /detect-bugs
│   │   │   ├── explain.py        ← POST /explain + POST /generate-tests
│   │   │   └── health.py         ← GET /health
│   │   ├── services/             ← Business logic (★ swap point for real ML)
│   │   │   ├── code_search.py    ← Mock semantic code search
│   │   │   ├── bug_detector.py   ← Mock bug detection
│   │   │   ├── llm_service.py    ← Mock code explanation + test generation
│   │   │   └── auth.py           ← Placeholder authentication
│   │   └── models/
│   │       └── schemas.py        ← Pydantic request/response contracts
│   ├── tests/
│   │   └── test_endpoints.py     ← Endpoint tests (pytest + httpx)
│   ├── requirements.txt
│   ├── Dockerfile
│   └── .env.example
│
├── notebooks/                    ← Jupyter Notebooks for ML model training
│   ├── A_code-search.ipynb       ← Code search embedding & indexing
│   ├── B_bug-detector.ipynb      ← Bug detection model training
│   └── C_finetune-codet5.ipynb   ← CodeT5+ fine-tuning
│
├── extension/                    ← VS Code extension (TypeScript)
│   ├── src/
│   │   ├── extension.ts          ← Entry point: activate() wires features
│   │   ├── hoverProvider.ts      ← Hover → POST /explain
│   │   ├── diagnostics.ts        ← On save → POST /detect-bugs
│   │   └── commands.ts           ← Command → POST /generate-tests
│   ├── package.json
│   ├── tsconfig.json
│   └── README.md
│
├── .github/workflows/
│   └── ci.yml                    ← GitHub Actions: pytest + tsc
│
├── docs/
│   ├── api-contracts.md          ← Exact JSON schemas for all endpoints
│   └── system-design.md          ← Architecture description
│
└── README.md                     ← This file

---

Running Tests

Backend Tests

cd backend
pip install anyio pytest-anyio    # async test support
pytest tests/ -v

Extension Compilation Check

cd extension
npm run compile

---

Docker

# Build
docker build -t code-assistant-backend ./backend

# Run
docker run -p 8000:8000 code-assistant-backend

---

Configuration

The extension's backend URL is configurable via VS Code settings:

Setting	Default	Description
codeAssistant.backendUrl	http://localhost:8000	Backend server URL
codeAssistant.enableHover	true	Enable/disable hover explanations
codeAssistant.enableDiagnostics	true	Enable/disable bug detection

---

Next Steps: Replace Mock Functions with Real Logic

The project is designed so that swapping mock → real is a single-file change per module. Each mock function in backend/app/services/ is marked with # TODO: replace with real model/API call.

What to swap

File	Function(s)	Replace with
services/code_search.py	search_code()	CodeBERT embeddings + FAISS/Pinecone vector search
services/bug_detector.py	detect_bugs()	Fine-tuned defect detection model or static analysis
services/llm_service.py	explain_code(), generate_tests()	OpenAI / Anthropic / Gemini / local LLM API calls
services/auth.py	verify_api_key()	Real API key or JWT validation

Rules for swapping

1. Keep the same function signatures — the API routers call these directly
2. Return the same types — use the Pydantic models from schemas.py
3. No changes needed in api/*.py, schemas.py, or the VS Code extension

See docs/api-contracts.md for the exact JSON shapes your implementations must match.

---

API Endpoints

Method	Path	Description
GET	/health	Health check
POST	/search	Semantic code search
POST	/detect-bugs	Bug / code smell detection
POST	/explain	Code explanation
POST	/generate-tests	Unit test generation

Full API docs available at http://localhost:8000/docs when the backend is running.

---

License

MIT