RediSearch NewsWire Demo

⚠️ Important: This demo uses relative API paths (/api) in the frontend so the UI and backend can be deployed together on the same host (VM, container, etc.).

Latency Expectations (Important)

Deployment	Typical Latency
Backend on Azure VM (same region as Redis)	~1.5–2 ms
Backend on local laptop → Azure Redis	~35–70 ms

Why this happens

When running on an Azure VM in the same region as Redis:

• Traffic stays inside Azure’s backbone network
• Very low RTT and consistent performance

When running from a local machine:

• Requests traverse the public internet
• TLS + WAN latency dominates total response time

✅ Best practice: For realistic performance testing and demos, run the backend in the same region as your Redis instance.

Prerequisites

1. Azure Managed Redis Instance

You need to create an Azure Managed Redis instance with the following specifications:

• SKU: Balanced B0/B1
• Features Required: RediSearch, RedisJson module enabled
• Region: Any (recommend same region as your location for better latency)

💡 Tip: For testing with 100 documents, B0/B1 provides comfortable headroom and good performance.

2. Python Environment

• Python 3.8 or higher
• pip package manager

3. Sample Data

• The repository includes a small sample parquet file (sample.parquet) with 100 documents
• You may substitute your own parquet file for larger datasets (10k+)

🚀 Quick Start

1. Clone the repository

git clone https://github.com/yourusername/redisearch-newswire-demo.git
cd redisearch-newswire-demo

2. Set up virtual environment

Mac/Linux:

python3 -m venv venv
source venv/bin/activate

Windows:

python -m venv venv
venv\Scripts\activate

3. Create environment file

Copy the template and add your credentials:

cp .env.template .env

Edit .env with your actual Azure Managed Redis credentials:

AMR_HOST=your-redis-instance.redis.cache.windows.net
AMR_PORT=10000
AMR_PASSWORD=your-redis-password-here

🔑 Where to find credentials:

1. Go to Azure Portal
2. Navigate to your Redis Cache instance
3. Go to "Access keys" section
4. Copy the hostname, port, and primary access key

4. Install dependencies

pip install -r requirements.txt

5. Create the RediSearch index

cd backend
python create_index.py

This creates the newswire_idx index with the proper schema for searching news articles.

6. Load sample data

python load_sample_data.py

This loads documents from your parquet file into Redis. The small sample takes ~30 seconds.

7. Start API server

python -m uvicorn main:app --host 0.0.0.0 --port 8000

8. Open in browser

On the VM:

http://127.0.0.1:8000

From your laptop:

http://<VM_PUBLIC_IP>:8000

Make sure port 8000 is allowed in your Azure VM network rules.

🧹 Cleanup

To delete all indexed data:

redis-cli FT.DROPINDEX newswire_idx DD