devchat-hm 0.1.1

DevChat – AI powered project-aware coding assistant

Copy🤖 DevChat - AI-Powered Local Coding Assistant

DevChat is an intelligent, project-aware coding assistant that understands your entire codebase using RAG (Retrieval-Augmented Generation) technology. Ask questions, get instant answers, and even fix bugs - all from your terminal!

🌟 Why DevChat? Traditional AI assistants don't understand your specific project. DevChat indexes your entire codebase, creates semantic embeddings, and gives you context-aware answers specific to YOUR code. ✨ Key Features

🧠 Project-Aware Intelligence - Understands your complete codebase context 🔍 Smart Code Search - Find functions, classes, and files using natural language 💬 Interactive Chat - Ask questions in plain English 🛠️ Automatic Bug Fixes - Detects and applies code fixes with approval ⚡ Lightning Fast - Uses FAISS vector search for instant results 🎯 Multi-Mode Operation - File-specific, project-wide, and general chat modes 📝 Smart File Detection - Automatically detects which files are relevant 🚀 Powered by Groq - Uses fast Llama 3.1 model for responses

📦 Installation Requirements

Python 3.9 or higher Groq API Key (Get one free here)

Install via pip bashpip install devchat-hm Set up your API key bash# Linux/Mac export GROQ_API_KEY="your-groq-api-key-here"

Windows (PowerShell)

$env:GROQ_API_KEY="your-groq-api-key-here"

Windows (CMD)

set GROQ_API_KEY=your-groq-api-key-here

🚀 Quick Start

1. Run DevChat bashdevchat
2. Enter your project path when prompted Enter project path: /path/to/your/project
3. Wait for indexing (one-time process) 🔍 Indexing project files... ✅ Indexed: /path/to/your/project/main.py ✅ Indexed: /path/to/your/project/utils.py ... 🤖 DevChat is ready. Ask questions.
4. Start asking questions! 🧑‍💻 > What does the main.py file do? 🤖 DevChat: The main.py file contains the entry point of the application...

🧑‍💻 > How is authentication handled? 🤖 DevChat: Authentication is implemented in auth.py using JWT tokens...

🧑‍💻 > Fix the bug in utils.py line 45 🤖 DevChat: FILE: utils.py START_LINE: 45 END_LINE: 48 CODE: def process_data(data): if data is None: return [] return [item.strip() for item in data]

Apply fix? (y/n):

🎯 How It Works DevChat operates in 3 intelligent modes: 1️⃣ File-Specific Mode When you mention a specific file (e.g., main.py, utils.js), DevChat retrieves only that file's content. 🧑‍💻 > What functions are in auth.py? 2️⃣ Project-Wide Mode For general project questions, DevChat searches across all indexed files using semantic similarity. 🧑‍💻 > Where is the database connection configured? 3️⃣ General Chat Mode For non-code questions, DevChat acts like a helpful AI assistant. 🧑‍💻 > Explain what REST APIs are

🛠️ Advanced Features Automatic Code Fixes DevChat can detect when you want to modify code and generates precise fixes: 🧑‍💻 > Fix the error in main.py line 23 - variable 'user' might be None 🤖 DevChat: FILE: main.py START_LINE: 23 END_LINE: 25 CODE: if user is not None: print(user.name) else: print("User not found")

Apply fix? (y/n): y ✅ Fix applied successfully! Supported Languages DevChat supports the following file types:

Python (.py) JavaScript (.js, .jsx) Java (.java) C/C++ (.c, .cpp, .h) HTML/CSS (.html, .css) JSON (.json) TypeScript (.ts)

Smart Folder Exclusion DevChat automatically ignores:

node_modules/ .git/ pycache/ dist/, build/ .next/ .venv/

🔧 Configuration DevChat uses a simple configuration system. Key settings are in devchat/config.py: pythonclass Config: GROQ_API_KEY = os.getenv("GROQ_API_KEY") MODEL = "llama-3.1-8b-instant" # LLM model EMBEDDING_MODEL = "all-MiniLM-L6-v2" # Embedding model VECTOR_DB_PATH = "devchat.index" # Index storage path Custom Configuration You can modify these settings by creating a custom config: pythonfrom devchat.config import Config

Config.MODEL = "llama-3.1-70b-versatile" # Use a larger model Config.VECTOR_DB_PATH = "custom_index.db" # Custom index location

📊 Technical Architecture ┌─────────────────────────────────────────────────────────┐ │ User Question │ └────────────────────┬────────────────────────────────────┘ │ ▼ ┌───────────────────────┐ │ Question Analyzer │ │ (File/Project/Chat) │ └───────────┬───────────┘ │ ┌────────────┴────────────┐ │ │ ▼ ▼ ┌──────────────┐ ┌──────────────┐ │ File Search │ │ Vector Store │ │ (Exact) │ │ (Semantic) │ └──────┬───────┘ └──────┬───────┘ │ │ └────────────┬────────────┘ │ ▼ ┌───────────────────────┐ │ Build Context │ │ (Top-K Documents) │ └───────────┬───────────┘ │ ▼ ┌───────────────────────┐ │ LLM (Groq/Llama) │ │ Generate Response │ └───────────┬───────────┘ │ ▼ ┌───────────────────────┐ │ Response Parser │ │ (Fix Detection) │ └───────────┬───────────┘ │ ▼ ┌──────────┐ │ Output │ └──────────┘ Core Components

Indexer (indexer.py) - Scans project and creates document embeddings Vector Store (vectorstore.py) - FAISS-based similarity search RAG (rag.py) - Builds context from relevant documents LLM (llm.py) - Groq API integration for responses Chat Loop (chat.py) - Main interaction logic with 3 modes Editor (editor/) - Code diff, parsing, and patching tools

💡 Usage Examples Example 1: Understanding Code Structure 🧑‍💻 > What classes are defined in models.py? 🤖 DevChat: The models.py file defines 3 classes:

1. User - Represents user accounts
2. Post - Represents blog posts
3. Comment - Represents post comments Example 2: Finding API Endpoints 🧑‍💻 > Where are the API routes defined? 🤖 DevChat: API routes are defined in routes/api.py:

• GET /api/users - Fetch all users
• POST /api/users - Create new user
• GET /api/posts - Fetch all posts Example 3: Debugging Errors 🧑‍💻 > Why am I getting "KeyError: 'user_id'" in app.py? 🤖 DevChat: In app.py line 34, you're accessing request.form['user_id'] without checking if it exists. Use request.form.get('user_id') instead to avoid KeyError. Example 4: Code Modification 🧑‍💻 > Update the database connection timeout to 30 seconds in config.py 🤖 DevChat: FILE: config.py START_LINE: 12 END_LINE: 12 CODE: DB_TIMEOUT = 30 # seconds

Apply fix? (y/n): y ✅ Fix applied successfully!

🎓 Best Practices

1. Be Specific ❌ "Tell me about the code" ✅ "What does the authenticate() function in auth.py do?"
2. Mention File Names Including file names helps DevChat give precise answers: 🧑‍💻 > How does error handling work in utils.py?
3. Use Action Keywords for Fixes Keywords like "fix", "update", "change", "modify" trigger fix mode: 🧑‍💻 > Fix the indentation error in main.py line 45
4. Ask Follow-up Questions DevChat maintains context within a session: 🧑‍💻 > What does the User class do? 🤖 DevChat: [explains User class]

🧑‍💻 > Show me the init method 🤖 DevChat: [shows init method]

🤝 Contributing Contributions are welcome! Here's how you can help: Reporting Issues Found a bug? Open an issue with:

Description of the problem Steps to reproduce Expected vs actual behavior Your Python version and OS

Feature Requests Have an idea? Open a feature request describing:

The feature you'd like Why it would be useful How it might work

Pull Requests

Fork the repository Create a feature branch (git checkout -b feature/amazing-feature) Make your changes Write/update tests if applicable Commit your changes (git commit -m 'Add amazing feature') Push to your fork (git push origin feature/amazing-feature) Open a Pull Request

📝 License This project is licensed under the MIT License - see the LICENSE file for details.

👨‍💻 Author Hemanth Meher

📧 Email: [e-mail] 🐙 GitHub: @hemanthmeher 💼 LinkedIn: Connect with me

🙏 Acknowledgments

Groq - Lightning-fast LLM inference FAISS - Efficient similarity search by Facebook AI Sentence Transformers - State-of-the-art text embeddings Click - Beautiful command-line interfaces

🔮 Roadmap Coming Soon

Support for more programming languages Integration with VS Code extension Multi-file code generation Project-wide refactoring suggestions Integration with GitHub for PR analysis Support for OpenAI GPT models Custom embedding models Web interface (optional GUI)

📊 Project Stats

Lines of Code: ~500 Dependencies: 6 core packages Supported File Types: 11+ Average Response Time: <2 seconds Index Size: Varies by project (typically 1-10 MB)

⚡ Performance Tips

1. Index Only What You Need If you have a large project, consider indexing specific directories: python# Modify indexer.py to limit scope project_path = "/your/project/src" # Instead of root
2. Use Faster Models For quicker responses, use smaller models: pythonConfig.MODEL = "llama-3.1-8b-instant" # Fast

vs

Config.MODEL = "llama-3.1-70b-versatile" # Accurate but slower 3. Reuse Indexes DevChat saves the index to devchat.index. On subsequent runs, it loads from disk instead of re-indexing (future feature).

❓ FAQ Q: Do I need an internet connection? A: Yes, DevChat uses the Groq API which requires internet. The vector search is local, but LLM inference is cloud-based. Q: Is my code sent to the cloud? A: Only the relevant code snippets matching your question are sent to Groq's API along with your prompt. Full codebase stays local. Q: Can I use OpenAI instead of Groq? A: Currently, DevChat is configured for Groq. OpenAI support is planned for future versions. Q: How accurate are the answers? A: DevChat is as accurate as the LLM (Llama 3.1) and the relevance of retrieved documents. It works best with well-structured codebases. Q: Can I use this in production? A: DevChat is great for development and debugging. For production AI features, consider rate limits and error handling. Q: How do I reset the index? A: Simply delete the devchat.index file and run DevChat again to re-index your project.

🚨 Troubleshooting Issue: "GROQ_API_KEY not found" Solution: Make sure you've set the environment variable: bashexport GROQ_API_KEY="your-key-here" Issue: No files being indexed Solution: Check if your files have supported extensions (.py, .js, etc.) and aren't in ignored folders. Issue: "NOT FOUND" responses Solution:

Ensure the file/code you're asking about actually exists in your indexed project Try rephrasing your question with more specific keywords Mention specific file names for better results

Issue: Slow indexing Solution:

Large projects take longer to index Exclude unnecessary directories by modifying IGNORE_DIRS in indexer.py