Qdrant Server Devcontainer For Rag MCP

🚀 Qdrant Development Container for File Embeddings

This project offers a development environment configuration for running Qdrant to perform file embeddings. It includes all necessary components to index and search text documents using vector similarity search technology.

🚀 Quick Start

Prerequisites

1. Docker Desktop must be running before starting the development container.
2. VS Code with the Remote - Containers extension installed.
3. Internet connection (for downloading dependencies).

Steps

1. Ensure Docker Desktop is running on your system.
2. Open this folder in VS Code.
3. Click the green "Reopen in Container" button in the bottom - right corner.
   • Or press F1 and type "Dev Containers: Reopen in Container".

📦 Project Structure

qdrant_server_devcontainer/
├── .devcontainer/
│   ├── devcontainer.json
│   └── Dockerfile
├── requirements.txt
├── ingest.py
└── data/ # Place your text files here

💻 Usage Examples

Basic Usage

1. Place your text files in the data/ directory.
2. The container will automatically start the Qdrant service.
3. After the container is built, you should be able to access Qdrant via http://localhost:6333.
4. Manually run the loading script from inside the container:

python ingest.py

✨ Features

• Qdrant vector database running in the background.
• Automatic file indexing using sentence - transformers.
• A Python environment equipped with all necessary dependencies.
• The VS Code Python extension pre - installed.

🔧 Technical Details

• Qdrant runs on a dynamically assigned port (check the container build output panel).
• Uses all - MiniLM - L6 - v2 for text embeddings.
• Creates a collection named "local - docs" using cosine similarity.
• Supports text files (.txt), Markdown files (.md), and PDF files (.pdf) in the data directory.

🛠️ Troubleshooting

Container fails to start

• Ensure Docker Desktop is running.
• Check if any other processes are occupying the dynamically assigned port.
• Verify that all dependencies are installed correctly.

Files are not indexed

• Check if the files are in the data/ directory.
• Confirm that the file extensions are supported (currently .txt, .md, .pdf).
• Ensure the files are readable by the container.

📄 License

MIT License

📋 To - Do List

• Handle large - volume PDF files.
• Use parallel processing to extract text by page.
• Embed and push each snippet as soon as it is ready.
• Add GPU embedding support if torch.cuda.is_available().
• Add support for epub files.