Skip to content

QDRANT_QUICK_START.md

Latest commit

 

History

History
163 lines (117 loc) · 3.44 KB

QDRANT_QUICK_START.md

File metadata and controls

163 lines (117 loc) · 3.44 KB

Qdrant Local Mode - Quick Start

✅ You're Already Using It!

Your project uses Qdrant in local/embedded mode - no server installation needed!

🎯 How It Works

Local Mode (Current Setup)

from qdrant_client import QdrantClient

# This line stores everything in ./qdrant_db/ directory
client = QdrantClient(path="./qdrant_db")

That's it! All vectors are stored as files in the qdrant_db/ folder.

📂 Where Is Data Stored?

document_search/
├── qdrant_db/              ← All your vectors stored here!
│   ├── meta.json
│   ├── collections/
│   │   └── documents/      ← Your collections
│   └── ...

To backup: Copy the qdrant_db/ folder
To reset: Delete the qdrant_db/ folder
To share: Zip and send the qdrant_db/ folder

🚀 Quick Examples

1. Basic Usage (Your Current Code)

from src.document_search import DocumentSearchRAG, Document

# Already configured for local mode!
rag = DocumentSearchRAG()

# Index documents (stored in ./qdrant_db/)
doc = Document(id="1", title="Test", content="Hello world")
rag.index_documents([doc])

# Search (reads from ./qdrant_db/)
results = rag.search("hello", top_k=5)

2. Direct Qdrant Client

from qdrant_client import QdrantClient

# Local mode
client = QdrantClient(path="./qdrant_db")

# List collections
collections = client.get_collections()
print(collections)

3. Check What's Stored

from qdrant_client import QdrantClient

client = QdrantClient(path="./qdrant_db")

# Get info about a collection
info = client.get_collection("documents")
print(f"Points stored: {info.points_count}")

🔄 Local vs Server Mode

Feature Local Mode ✅ (Current) Server Mode
Installation None needed Docker/Binary
Storage Files in directory Server database
Performance Good for dev Better for production
UI Dashboard No Yes (http://localhost:6333/dashboard)
Multiple Clients No Yes
Best For Development, Testing Production, Large Scale

💡 Common Operations

View Collections

source .venv/bin/activate
python -c "
from qdrant_client import QdrantClient
client = QdrantClient(path='./qdrant_db')
for col in client.get_collections().collections:
    print(col.name)
"

Get Collection Stats

python -c "
from qdrant_client import QdrantClient
client = QdrantClient(path='./qdrant_db')
info = client.get_collection('documents')
print(f'Points: {info.points_count}')
print(f'Vector size: {info.config.params.vectors.size}')
"

Clear a Collection

python -c "
from qdrant_client import QdrantClient
client = QdrantClient(path='./qdrant_db')
client.delete_collection('documents')
print('Collection deleted!')
"

📚 Run the Example

source .venv/bin/activate
python examples/use_local_qdrant.py

This will show you:

  • How local mode works
  • How to index and search
  • How to inspect storage
  • Comparison with server mode

🎉 Bottom Line

You don't need to install anything!

Qdrant local mode is already working in your project. Just use your RAG classes and everything is stored in ./qdrant_db/.

Want to try it? Run:

source .venv/bin/activate
python quick_test.py

This will test your RAG system with local Qdrant storage!